unslothai/unsloth

What it measures: Copy-pasted code that should be shared instead.

Method: Code duplication via token-stream sliding windows with type-aware normalization (locals masked, type names preserved), density-scored per KLoC of production code. Deterministic.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Verified

30 duplicated block group(s) detected. A further 3 rows report members as variants of one another; they aggregate block groups already counted above and are not themselves counted. Measured on part of this repository only: .py (60% of production source) was not exposed to the token comparison, so duplication there is unmeasured and is not in this count. Measured on part of this repository only: .ts, .tsx (37% of production source) was not exposed to THIS dimension's token comparison and is not in this count; duplication there is measured by R10 Code Duplication, the frontend lens's card running the same clone algorithm over the JS/TS token stream.

+ 21 more group(s) — more in Appendix A; the complete list is findings.md.

What to do

1. Resolve the 3 Duplicated block (12 lines × 2) finding(s) in Code Duplication — start with process.rs, desktop_update_policy.rs, backend.rs. — One of this dimension's main actionable groups (3 warning-level).
2. Resolve the 3 Duplicated block (9 lines × 2) finding(s) in Code Duplication — start with desktop_update_policy.rs (2), phase_log.rs. — One of this dimension's main actionable groups (3 warning-level).
3. Resolve the 3 Duplicated block (7 lines × 2) finding(s) in Code Duplication — start with install.rs (2), desktop_backend_owner.rs. — One of this dimension's main actionable groups (3 warning-level).
4. Enforce Code Duplication in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.

Detailed fixes: d4_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether volatile projects sit underneath others that depend on them (so their churn ripples upward), and whether project dependencies form cycles. A widely-depended-on but stable shared/kernel project is healthy, not penalised.

Method: Dependency cycles via elementary-DFS over real .csproj references, plus Martin instability (afferent/efferent) per project. Exhaustive over the reference graph, deterministic.

Coverage: Exhaustive · type-level: afferent/efferent coupling + cycles computed over every production type — the population is all types, not a name convention.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Prevented

7 production modules (Cargo+npm+Python), 0 dependency cycle(s), 0 unstable depended-on module(s). Read from the build's own module declarations; 0 module(s) off the main sequence, with abstractness counted on 4 of the 7 (the rest declare no modelled class or interface, export only macros, declare types in source no model reads, or have no source directory of their own).

✓ On the Gold path — maintain.

Detailed fixes: d5_recommendation.md.

What it measures: Whether a class's methods are focused on a single responsibility.

Method: LCOM4 cohesion per production class with at least two methods: connected components of methods sharing state or calls, computed syntactically. Deterministic, not a proxy.

Coverage: Exhaustive · type-level: LCOM4 cohesion computed over every production class — the population is all types, not a name convention.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Verified

0 of 44 classes have LCOM4 above 3.

✓ On the Gold path — maintain.

Detailed fixes: d6_recommendation.md.

What it measures: Whether the test suite has a healthy mix of unit / integration / end-to-end tests.

Method: Test projects classified (Unit/Integration/BDD/E2E) from compiled metadata; test methods counted exhaustively across projects with placement-agnostic disk fallback. Deterministic.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Documented

84577 test methods: 84494 unit, 50 integration, 0 BDD, 33 e2e. The JavaScript/TypeScript suite contributes 12824 `it`/`test` case(s) across 1369 test file(s) declaring at least one; its tier split is read from package names and paths only. The Rust suite contributes 818 `#[test]` function(s) across 41 file(s) declaring at least one; its unit/integration split is Cargo's own — 0 of those file(s) are integration-test targets under a crate's tests/ directory, and the rest are #[test] functions compiled into the crate they test. The Python suite contributes 70935 test function(s) across 3300 file(s) declaring at least one — every `def test…` in a file pytest or unittest would collect, which is those frameworks' own definition of a case; a parametrize table counts once, so this is a floor. Its tier split is read from file names and paths only.

✓ On the Gold path — maintain.

Detailed fixes: d9_recommendation.md.

What it measures: Whether the tests truly assert behaviour rather than just running the code.

Method: Per-test assertions, skips, and mock references analyzed via Roslyn; structured skip-reason tags (BUG:/ENV:) separate documented deferrals from debt. Deterministic.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Prevented

13 skipped (13 with a documented reason), 12 zero-assertion, 12 synchronised by a fixed sleep, no mocking-framework packages referenced (hand-written doubles or no mocking) across 12824 tests. Measured on the JavaScript/TypeScript suite only — at least 3,570 test source file(s) (.py, .ps1) went unread, so its test quality is unmeasured and is not in these counts.

What to do

1. Resolve the 12 No assertions finding(s) in Test Quality — start with think-tag-tracker.test.ts (6), stuck-generation-run.test.ts, sidebar-pin-without-matchmedia.test.ts. — One of this dimension's main actionable groups (12 warning-level).
2. Resolve the 12 Fixed-sleep synchronisation finding(s) in Test Quality — start with recipe-server-persistence.test.ts (2), qwen-defaults-migration-safety.test.ts (2), sidebar-server-paging.test.ts. — One of this dimension's main actionable groups (12 warning-level).
3. Enforce Test Quality in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.

Detailed fixes: d10_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether dependencies are current, secure, and not bloated.

Method: Manifest scan via dotnet list package across all projects; worst-signal-per-package deduction (saturating for vulnerabilities, capped-linear for deprecation/outdated) per KLoC. Exhaustive, deterministic.

Maturity: Documented → Verified → Prevented · effective 9.8 / 10 · rule-coverage 100% · ceiling Verified

4 outdated, 0 yanked direct Cargo dependencies. Only DIRECT edges are graded: a transitive crate cannot be moved past what its parent's requirement admits, so reporting one would be advice its owner cannot take. A newer release is reported only where this repository's OWN requirement already admits it, so the remedy is `cargo update` and never a manifest edit — which means a release outside the declared range is deliberately NOT charged, because a written-down constraint is a decision rather than a defect. Note that a bare requirement is a CARET, and for a 0.x crate its ceiling is the minor. Whether any crate is UNMAINTAINED is not graded — crates.io publishes no maintenance status, and release age does not stand in for one. Known CVEs in this dependency graph are D30's question, read from REDACTED there.

✓ On the Gold path — maintain.

Detailed fixes: d12_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether any secrets (keys, tokens, passwords) have leaked into the code.

Method: In-process native secret scanner (entropy plus signature patterns) across all tracked files; no external tool. A clean result is a measured 10, not no-data zero. Deterministic.

Maturity: Documented → Verified → Prevented · effective 0.0 / 10 · rule-coverage 100% · ceiling Prevented

4 secret(s) detected.

What to do

1. Resolve the 4 Leaked secret finding(s) in REDACTED Scanning — start with REDACTED, REDACTED, REDACTED. — One of this dimension's main actionable groups (4 issue-level).
2. Enforce REDACTED Scanning in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.

Detailed fixes: d13_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether the licenses of third-party packages are compatible with your policy.

Method: Third-party package licenses resolved from declared package metadata and checked against the configured policy (allow/deny/copyleft). Deterministic; clean = no incompatible license found at metadata depth.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Verified

0 of 615 shipped crate(s) use a banned license. Licences were resolved from crates.io over the crates a consumer compiles — this repository's REDACTED closed over its manifests' `[dependencies]` and `[build-dependencies]`. Crates it asks for ONLY under `[dev-dependencies]` are excluded: they are not compiled by anything that depends on this repository. ★ COVERAGE OF THIS VERDICT: it grades this repository's crate dependencies and nothing else. The repository also declares a Python pyproject.toml/requirements.txt (pip/uv/Poetry) and package.json, and the licences of those dependency graphs were NOT read by this pass — a gap in this engine's coverage, not a statement about them. So this result says the graded closure carries no banned licence; it does NOT say this repository's licensing is clear.

✓ On the Gold path — maintain.

Detailed fixes: d14_recommendation.md.

What it measures: Whether knowledge is concentrated in too few people (the "bus factor").

Method: Living knowledge per author via time-decayed commit attribution (6-month half-life, focus weighting) across largest source files. Deterministic, avoids blame's mechanical-refactor false positives.

Maturity: Documented → Verified → Prevented · effective 9.0 / 10 · rule-coverage 100% · ceiling Documented

231 source file(s) have their living knowledge concentrated in one author (≥90% of recent, decayed contribution). The largest is REDACTED. Counted over 2437 of the 3383 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.

✓ On the Gold path — maintain.

Detailed fixes: d16_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Acknowledged debt left in the code — TODOs, dead code, suppressed warnings.

Method: Roslyn syntactic debt markers (suppressions/TODO/FIXME/HACK/empty-catch/commented-code/Obsolete) plus SymbolFinder dead-code analysis; weighted-debt-per-KLoC density deducted 2.0x per unit. Deterministic, exhaustive.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Prevented

22 deducted task-comment markers across 1465212 LoC (0.0/KLoC) → score 10.0. Task comments only: this repository's language is read without a compiler, so D17's suppression, dead-code and commented-out-code arms did not run and this score counts fewer marker kinds than a .NET repository's would.

What to do

1. Resolve the 21 TodoComment finding(s) in Explicit Debt — start with loader.py (6), model_config.py (2), REDACTED (2). — One of this dimension's main actionable groups (21 warning-level).
2. Resolve the 1 HackComment finding(s) in Explicit Debt — start with REDACTED. — One of this dimension's main actionable groups (1 warning-level).
3. Enforce Explicit Debt in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.

Detailed fixes: d17_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether the project's documentation is clear, complete, and useful.

Method: Judged by language model at low temperature (0.0-0.1) on a deterministic doc sample (READMEs plus first 25 architecture docs), with two-pass stability filtering. Advisory, sampled.

Maturity: Documented → Verified → Prevented · effective Strong / 10 · rule-coverage 100% · ceiling Documented

Both READMEs are per-directory references documents (t4_smoke/references and studiobench/symbols) rather than the repository root, so they each document their own directories. The t4_smoke one is a committed-metric-references guide with an overview, which leg applies to, and what it must not; the studiobench one is a symbol-bridge description plus a table of fields and failure-mode rationale. Both are well written for their niche. (10 of 12 sampled documents could not be assessed: 1 of 2 evaluation groups failed.)

What to do

1. Improve Documentation Quality — currently 8.0/10. — Both READMEs are per-directory references documents (t4_smoke/references and studiobench/symbols) rather than the repository root, so they each document their own directories. The t4_smoke one is a committed-metric-references guide with an overview, which leg applies to, and what it must not; the studiobench one is a symbol-bridge description plus a table of fields and failure-mode rationale. Both are well written for their niche. (10 of 12 sampled documents could not be assessed: 1 of 2 evaluation groups failed.)

Detailed fixes: d19_recommendation.md.

What it measures: Whether architecture decisions are recorded well (context, decision, consequences).

Method: Per-ADR judgment by language model at low temperature with two-pass stability; confidence is share of ADRs evaluated; enforcement-field presence detected deterministically. Advisory.

Maturity: Documented → Verified → Prevented · effective 0.0 / 10 · rule-coverage 100% · ceiling Documented

No architecture decision records were found.

What to do

1. Resolve the 1 No ADRs found finding(s) in ADR Quality. — One of this dimension's main actionable groups (1 recommendation-level).

Detailed fixes: d20_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether names — types, methods, variables — are clear and consistent.

Method: Judged by language model at low temperature (0.0-0.1) on a deterministic random symbol sample (fixed size, not exhaustive), with disclosed confidence band. Advisory, sampled.

Maturity: Documented → Verified → Prevented · effective Exemplary / 10 · rule-coverage 100% · ceiling Verified

0 naming inconsistencies across 0 sampled symbols.

✓ On the Gold path — maintain.

Detailed fixes: d21_recommendation.md.

What it measures: Whether any secrets were ever committed — scanned across the full git history, not just now.

Method: REDACTED scan via TWO gitleaks detect passes in an isolated checkout — the full git history, then a second --no-git pass over the working tree as it stands — merged and de-duplicated by (rule, file, line); each match flagged High. Both invocations are recorded in the audit trail. Exhaustive; when the tool is absent, or when its output cannot be parsed into the expected shape, the dimension is WITHHELD as an explicit measurement gap on our side — unscored and excluded from the lens, never a hedged middling score.

Maturity: Documented → Verified → Prevented · effective 7.0 / 10 · rule-coverage 100% · ceiling Documented

3 finding(s): 0 critical, 3 high, 0 medium, 0 low. Remediation for historically-committed secrets is credential rotation — they remain in history regardless of later deletion.

What to do

1. Resolve the 3 REDACTED finding(s) in Secrets (history) — start with REDACTED, REDACTED, REDACTED. — One of this dimension's main actionable groups (3 issue-level).
2. Resolve the 1 Rotate the exposed credentials finding(s) in Secrets (history). — One of this dimension's main actionable groups (1 recommendation-level).

Detailed fixes: d28_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Real static-analysis (SAST) findings — likely security bugs in the code, any language.

Method: Polyglot static analysis via semgrep across the repo using the pinned, image-baked p/security-audit + p/owasp-top-ten rulesets (no scan-time registry fetch); severity rules (ERROR/WARNING/INFO) map to a full-band severity-weighted score. Exhaustive, deterministic; degrades on parse failure.

Coverage: semgrep pattern rules over all files — exhaustive for the rule set, blind to classes of bug without a rule (clean = no rule matched).

Maturity: Documented → Verified → Prevented · effective 0.0 / 10 · rule-coverage 100% · ceiling Documented

285 finding(s): 0 critical, 162 high, 91 medium, 32 low. semgrep hit a parse error in 115 file(s) — `.github/scripts/agent-guides-drive.sh` (line 25, line 26, lines 532–538), `.github/scripts/agent-guides-install.sh` (line 14), `.github/scripts/assert-llama-loads.sh`, `.github/scripts/assert-prompt-cache.sh` (line 47, line 152, line 203), `.github/scripts/clean-machine-assert.sh`, … (+110 more) — so no absence of findings in the named regions is evidence of anything; rows reported elsewhere in those files are real. Fix the syntax error (or exclude the file deliberately) and re-scan to cover them. Separately, one or more rules could not re-parse an embedded snippet in 42 file(s) (e.g. a workflow `run:` block read as shell). Those files WERE scanned and their other rows are unaffected; only those rules' view of those snippets is missing.

+ 27 more group(s) — more in Appendix A; the complete list is findings.md.

What to do

1. Resolve the 54 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (7), REDACTED (4), REDACTED (4). — One of this dimension's main actionable groups (54 issue-level).
2. Resolve the 48 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (13), REDACTED (9), REDACTED (6). — One of this dimension's main actionable groups (48 issue-level).
3. Resolve the 26 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (9), REDACTED (3), REDACTED (2). — One of this dimension's main actionable groups (26 issue-level).

Detailed fixes: d29_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether any dependency has a known published vulnerability (CVE), direct or transitive, in ANY ecosystem the repository declares — Dart pub, Elixir and Erlang via Hex, Go modules, Java and Kotlin via Maven/Gradle, JavaScript/npm, .NET/NuGet, PHP/Composer, Python/PyPI, RubyGems, Rust/Cargo and Swift.

Method: Dependency-CVE scan across every ecosystem the repository declares, scored ONCE. Three sources are unioned and deduplicated by advisory identity (rule id + alias closure, CVE<->GHSA) scoped to package+version, keeping the worst severity: `osv-scanner --recursive` over osv.dev for Dart pub, Elixir/Hex (and Erlang, whose `rebar.lock` syft first converts to a CycloneDX SBOM the scanner reads, with rows attributed back to the lock), Go, Java and Kotlin via Maven/Gradle (and Scala, whose sbt build's pinned direct declarations are written into a CycloneDX SBOM the scanner reads, with rows attributed back to the build file), npm, PHP/Composer, Python/PyPI, RubyGems, Rust/Cargo and Swift; `trivy fs --scanners vuln` for npm lockfiles; and `dotnet list package --vulnerable --include-transitive` for NuGet (with per-advisory collapse of the project x target-framework fan-out), plus a DECLARED-dependency arm that resolves a published gem's gemspec against rubygems.org where no Gemfile.lock is committed. `SeverityScore(c,h,m,l, normalizer 8.0)`. NotApplicable only when NO ecosystem is readable; if any applicable ecosystem could not be scanned the findings are REPORTED and the score is withheld. Supersedes the npm and OSV arms, retired 2026-09-05.

Maturity: Documented → Verified → Prevented · effective 5.4 / 10 · rule-coverage 100% · ceiling Documented

16 finding(s): 0 critical, 4 high, 11 medium, 1 low.

What to do

1. Resolve the 4 High CVE finding(s) in Dependency Vulnerabilities — start with REDACTED (4). — One of this dimension's main actionable groups (4 issue-level).
2. Resolve the 4 Medium advisory (unmaintained) finding(s) in Dependency Vulnerabilities — start with REDACTED (4). — One of this dimension's main actionable groups (4 warning-level).
3. Resolve the 4 Medium CVE finding(s) in Dependency Vulnerabilities — start with REDACTED (2), REDACTED (2). — One of this dimension's main actionable groups (4 warning-level).

Detailed fixes: d30_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether Dockerfiles / Terraform / Kubernetes config follow security best practices.

Method: IaC/container misconfiguration scan via trivy config (REDACTED/Terraform/K8s/Helm/CloudFormation); severity rules to 0-10 moderate normalizer. NotApplicable without manifests. Exhaustive, deterministic.

Maturity: Documented → Verified → Prevented · effective 5.9 / 10 · rule-coverage 100% · ceiling Documented

29 finding(s): 0 critical, 6 high, 20 medium, 3 low.

What to do

1. Resolve the 20 Medium IaC finding(s) in IaC & Container Security — start with REDACTED (6), REDACTED (5), REDACTED (5). — One of this dimension's main actionable groups (20 warning-level).
2. Resolve the 6 High IaC finding(s) in IaC & Container Security — start with REDACTED (2), REDACTED (2), REDACTED. — One of this dimension's main actionable groups (6 issue-level).
3. Resolve the 3 Low IaC finding(s) in IaC & Container Security — start with REDACTED, REDACTED, REDACTED. — One of this dimension's main actionable groups (3 recommendation-level).

Detailed fixes: d31_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether anyone still has living knowledge of each file, or it has been orphaned — last understood long ago by someone now gone quiet. The sibling of the bus factor: D16 asks who owns it, D34 asks whether anyone still knows it.

Method: File orphaning as total living-knowledge decay below one focused-commit's worth within a year, computed per-file from the D16 decay model. Exhaustive, deterministic over fixed history.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Documented

Every significant source file has living knowledge — recently and meaningfully worked. Counted over 2437 of the 3383 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.

✓ On the Gold path — maintain.

Detailed fixes: d34_recommendation.md.

What it measures: Whether files that change together actually belong together — pairs that repeatedly co-change in git history despite having no explicit code dependency, surfacing the hidden/logical coupling (and boundaries in the wrong place) a static scan can't see.

Method: Pairwise co-occurrence over the per-commit file sets in git history (production source only — tests and generated dropped): Degree-of-Coupling = shared ÷ min individual revisions, reported above noise floors (each file ≥10 revisions, ≥5 shared commits, ≥50% strength); sweeping commits excluded. Deterministic over fixed history.

Coverage: Population: PRODUCTION source files only — test and generated files are dropped before pairing, so a class co-changing with its own test (trivially ~100%) can't drown the real production↔production coupling. Pairs ranked by Degree-of-Coupling. A non-source file is never a coupling PARTICIPANT either: documentation, schemas, config and data files are dropped with the rest, so a code↔docs pair — a command and the reference page that restates it — is not reported however strongly the two co-change; nor is coupling that runs THROUGH a build step or config file.

Maturity: Documented → Verified → Prevented · effective 9.6 / 10 · rule-coverage 100% · ceiling Documented

Strongest change-coupling: REDACTED↔REDACTED 100%; tool-ui-python.tsx↔tool-ui-terminal.tsx 100%; image_gallery.py↔api.ts 93%

What to do

1. Resolve the 38 Change coupling finding(s) in Change Coupling — start with index.ts (3), thread.tsx (2), training.py (2). — One of this dimension's main actionable groups (38 warning-level).
2. Resolve the 7 Change-coupling hub finding(s) in Change Coupling — start with REDACTED, index.ts, desktop_auth.rs. — One of this dimension's main actionable groups (7 warning-level).
3. Resolve the 1 Change coupling clique finding(s) in Change Coupling — start with tool-ui-code-execution.tsx. — One of this dimension's main actionable groups (1 warning-level).

Detailed fixes: d35_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether the build pipeline provides supply-chain integrity — generated provenance/attestation, signed artifacts (cosign/sigstore), an SBOM, and pinned build actions. Presence of the configuration, not a runtime guarantee.

Method: Supply-chain provenance/signing read deterministically from CI/build config (.github/workflows, .gitlab-ci.yml, azure-pipelines, Jenkinsfile, .circleci) + the release surface: four signals — generated provenance/attestation (SLSA/in-toto/actions-attest), artifact signing (cosign/sigstore/gitsign), an SBOM (syft/sbom-action/*.spdx.json/*.cdx.json), and SHA-pinned build actions — scored 10·present/denom. NotApplicable without a build pipeline. Detects configuration presence, not runtime enforcement.

Maturity: Documented → Verified → Prevented · effective 7.5 / 10 · rule-coverage 100% · ceiling Documented

3/4 supply-chain integrity signals present (provenance, signing, SBOM, pinned actions).

What to do

1. Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 warning-level).
2. Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 warning-level).
3. Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 warning-level).

Detailed fixes: d36_recommendation.md · top locations in Appendix A, every location in findings.md.

What it measures: Whether any dependency the repository declares is published as MALICIOUS rather than merely vulnerable — a package that is an attacker's work, in any ecosystem osv-scanner reads. Scored apart from D30 because the answer is binary: there is no safe version to upgrade to, and the fix is to remove the package and rotate every credential it could have read.

Method: The same dependency scan D30 reads, partitioned on the scanner's own classification rather than rescanned: a row is MALICIOUS when its id is in the `MAL-` space (the ossf/malicious-packages feed) OR its `database_specific.cwe_ids` carries `CWE-506` ("Embedded Malicious Code"). Both channels are structural; the summary text is deliberately NOT read, because a malicious-package record whose summary says only "Critical severity vulnerability" is a real shape ([GHSA redacted]) and a text matcher misses it. Scored BINARY: any surviving row is 0, whatever its severity and however many CVEs sit beside it — a hostile dependency is not a quantity. Applicability and degradation are D30's: NotApplicable only when no ecosystem is readable, and an unscannable ecosystem degrades rather than reading clean. SCORED, not informational.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Documented

No dependency in any ecosystem this repository declares is published as malicious.

✓ On the Gold path — maintain.

Detailed fixes: d43_recommendation.md.

What it measures: Whether anyone still ships security patches for the platform this repository RUNS ON — the runtime it pins and the framework majors its own constraints hold it to. Separate from D12 because the question differs: a current Django on an end-of-life Python is perfectly up to date and completely unsupported, and the fix is a migration rather than a version bump. What the repository says it merely SUPPORTS is never charged.

Method: End-of-life PLATFORM read from the repository's own declarations and graded against a FROZEN, dated table of vendor support dates — no network, no feed, no API, so this dimension answers identically inside a closed scan fence. Two subjects: a RUNTIME the project pins (a single or all-end-of-life TargetFramework, a .nvmrc or .python-version, a requires-python CAP) and a FRAMEWORK major a dependency constraint cannot move off (a caret, tilde or exact version; `vue@^2.7.16` pins Vue 2). A FLOOR is deliberately never charged — `requires-python = ">=3.8"` states what a package SUPPORTS, not what it runs on — and a multi-target project is charged only when EVERY target is out of support. Runtime 4.0/product capped 8.0, framework 1.5 capped 4.5. The table is safe to freeze because a statement about support that ended in the past cannot become false: it loses recall as it ages, never precision, and a test asserts every entry predates the freeze date. Disjoint from D31 (a container image's OS layer) and D29 (the toolchain a CI workflow installs). Abstains when the repository declares no platform this pass reads — never scores it clean.

Maturity: Documented → Verified → Prevented · effective 10.0 / 10 · rule-coverage 100% · ceiling Documented

0 end-of-life runtime(s) and 0 end-of-life framework(s), read from 1 platform declaration(s) and 19 dependency declaration(s). This dimension reads what the repository says about ITSELF — a pinned target framework, a version file, a capped requires-python, a Rust toolchain pin, a framework major a constraint cannot move off. A FLOOR is deliberately never charged: `requires-python = ">=3.8"` states what the package SUPPORTS, not what it runs on, and a well-maintained library declares exactly that while running its own CI on a current release. The end-of-life facts are FROZEN and dated, so this dimension needs no network and answers identically inside a closed scan fence; as the table ages it loses recall and never precision, because a statement about support that ended in the past cannot become false. The OS layer of a container image is D31's question and the toolchain a CI workflow installs is D29's; this row is neither.

✓ On the Gold path — maintain.

Detailed fixes: d44_recommendation.md.

Other · Accessibility — Whether non-text content carries a text alternative — img/area/input[type=image] have alt, a meaningful svg has a title or aria-label, video has a captions track, and object/embed/canvas have a name or fallback content. Static markup readiness, not a WCAG conformance claim.

Method: Static markup-model scan: every img/area/input[type=image] checked for alt, svg[role=img] for a title/aria-label, video for a captions <track>. Components skipped, spreads suppressed. Deterministic, hard fact per element.

Coverage: Population: image/media elements — img, area, input[type=image], svg, video, object, embed, canvas — across the PARSED MARKUP files only (.html/.htm/.cshtml/.razor/.vue/.svelte/.jsx/.tsx); components, hidden subtrees and dynamic-attribute elements are skipped. Markup built in script — tagged-template (html`…`) UIs and hyperscript DOM factories — is NOT read by any producer, so it contributes no element to this population; where such a frontend is present the card discloses it as an analyzer gap rather than scoring around it.

• Video with no captions <track> offers no captions. This player's source is set at runtime, so there is no caption file to author here: give the player a way to carry captions — accept a track/captions URL alongside the media and render a <track kind="captions"> when one is supplied — and capture that asset wherever the media enters the product (upload, import or generation). (×7) — studio/frontend/src/components/assistant-ui/attachment-preview.tsx:664, studio/frontend/src/components/assistant-ui/tool-ui-image-generation.tsx:237, studio/frontend/src/features/browser/video-file.tsx:162, …
• A <canvas> with no aria-label/title and no inner fallback content can't be described by assistive tech. Add a name or fallback content. (×3) — studio/frontend/src/components/file-viewer/pdf-view.tsx:87, studio/frontend/src/features/images/images-page.tsx:1209, studio/frontend/src/features/images/localized-edit-canvas.tsx:223

What to do

• Add a text alternative — alt on images (alt="" for purely decorative ones), a title or aria-label on meaningful svg, an aria-label or inner fallback content on canvas, and a captions <track> on video.

Other · Accessibility — Whether form controls have a programmatic label (an associated label, aria-label or aria-labelledby), buttons have text, links have an accessible name, a click handler on a plain element names the control it declares, fieldsets have a non-empty legend, known UI-library field components carry a label prop, and a placeholder isn't used as the only label. Static markup readiness, not a WCAG conformance claim.

Method: Static markup-model scan: inputs/selects/textareas checked for an associated label[for]/wrapping label/aria-label/aria-labelledby (per document), buttons for accessible text, fieldsets for a legend; placeholder-only labelling flagged. Deterministic, hard fact per control.

Coverage: Population: form controls, buttons, links, fieldsets and known UI-library field components in the PARSED MARKUP files only (.html/.htm/.cshtml/.razor/.vue/.svelte/.jsx/.tsx); components, hidden subtrees and spread/dynamic-attribute elements are skipped, so a control whose label arrives through a spread or a runtime expression is deliberately not judged. Markup built in script — tagged-template (html`…`) UIs and hyperscript DOM factories — is not read at all.

• This control has only a placeholder — a placeholder is not a label (it vanishes on input and many AT ignore it). Add a <label for>, a wrapping <label>, or aria-label. — studio/backend/assets/preview_page.html:291
• This control has only a placeholder — a placeholder is not a label (it vanishes on input and many AT ignore it). Add a <label htmlFor>, a wrapping <label>, or aria-label. (×9) — studio/frontend/smoke-find-in-page-main.tsx:319, studio/frontend/src/features/chat/prompt-storage/prompt-storage-dialog.tsx:1703, studio/frontend/src/features/chat/prompt-storage/prompt-storage-dialog.tsx:1715, …
• This control has no associated label. Add a <label htmlFor> / wrapping <label> / aria-label / aria-labelledby so assistive tech can name it. (×3) — studio/frontend/smoke-nonmodal-menus-main.tsx:211, studio/frontend/src/components/assistant-ui/audio-player.tsx:107, studio/frontend/src/components/tauri/diagnostics-copy-actions.tsx:63
• This <article> carries a click handler, which makes it a control, but it has no text, aria-label or labelled child (the icon-only case), so assistive tech has no name to announce for it. Note that switching it to <button type="button"> — the right fix for its keyboard operability — does not give it a name either: an icon-only button still needs one. Add visible text or an aria-label. — REDACTED:159
• This UI-library field component is named only by a placeholder — a placeholder is not a label (it vanishes as soon as the user types, and many assistive technologies ignore it). Give it a real name whichever way this library supports: a label prop, an aria-label, or an id on the rendered control with a <label htmlFor> pointing at it. For a group of controls, name the group itself (aria-label, or a fieldset with a legend) — labelling each item leaves the set unnamed. (×7) — studio/frontend/src/components/negative-prompt-field.tsx:66, studio/frontend/src/features/chat/chat-settings-sheet.tsx:1861, studio/frontend/src/features/chat/projects-page.tsx:1228, …
• This <video> carries a click handler, which makes it a control, but it has no text, aria-label or labelled child (the icon-only case), so assistive tech has no name to announce for it. Note that switching it to <button type="button"> — the right fix for its keyboard operability — does not give it a name either: an icon-only button still needs one. Add visible text or an aria-label. — studio/frontend/src/features/browser/video-file.tsx:162
• This <label> has no htmlFor and wraps no form control, so it names nothing: assistive tech never announces it, and clicking the caption focuses no field. Note that an id on the label is not an association — it is the TARGET of one, so a control still has to point at it. Give the label htmlFor="<the control's id>", move the control inside the label, or point the control's aria-labelledby at this label's id. (×10) — studio/frontend/src/features/export/components/export-run-panel.tsx:283, studio/frontend/src/features/export/components/export-run-panel.tsx:341, studio/frontend/src/features/export/components/export-run-panel.tsx:351, …
• A button with no text and no aria-label has no accessible name. Add visible text or an aria-label (an icon-only button still needs one). (×4) — studio/frontend/src/features/export/components/method-picker.tsx:34, studio/frontend/src/features/export/components/quant-picker.tsx:51, studio/frontend/src/features/export/export-page.tsx:1205, …
• This UI-library field component has no label / aria-label / aria-labelledby / id, and nothing it renders carries one either (a name prop is the form-submission key, not a label). The text inside it does not stand in for one: these compound APIs render a trigger with role="combobox" (or role="button"), and those roles take their name from the AUTHOR, not from their content — so the words on screen never reach the accessibility tree as a name. Name it whichever way this library supports: a label prop, an aria-label, or an id on the rendered control with a <label htmlFor> pointing at it. For a group of controls, name the group itself (aria-label, or a fieldset with a legend) — labelling each item leaves the set unnamed. (×2) — studio/frontend/src/features/export/export-page.tsx:1934, studio/frontend/src/features/export/export-page.tsx:2036
• This UI-library field component has no label / aria-label / aria-labelledby / id — and neither does anything it renders (a name prop is the form-submission key, not a label) — so it likely renders an unlabelled control. Name it whichever way this library supports: a label prop, an aria-label, or an id on the rendered control with a <label htmlFor> pointing at it. For a group of controls, name the group itself (aria-label, or a fieldset with a legend) — labelling each item leaves the set unnamed. (×2) — studio/frontend/src/features/images/hires-controls.tsx:119, studio/frontend/src/features/images/hires-controls.tsx:137

What to do

• Give every control a programmatic label (a <label for> / wrapping <label> / aria-label) and every button text — a placeholder is not a label.

Other · Accessibility — Whether pages declare a language (well-formed BCP-47) and a non-empty title, expose exactly one main landmark and a sane heading order with non-empty headings, keep zoom enabled, title their iframes, give data tables header cells, and avoid meta-refresh. Static markup readiness, not a WCAG conformance claim.

Method: Static markup-model scan: html lang, document <title>, a main landmark and heading order on full documents only, plus zoom-disabling viewports, untitled iframes and meta-refresh anywhere. Deterministic, per structural checkpoint.

Coverage: Population: the PARSED MARKUP documents (.html/.htm/.cshtml/.razor/.vue/.svelte/.jsx/.tsx). The page-level checks — lang, title, single main landmark — fire ONCE PER FULL DOCUMENT (an <html> root) and never on a partial or component fragment, so a repo of fragments is assessed only on the per-element checks (heading order, table headers, iframe titles, meta-refresh, zoom). Markup built in script — tagged-template (html`…`) UIs and hyperscript DOM factories — is not read at all.

• A <table> with data cells but no <th> gives assistive technology no way to associate cells with their headers. Mark the header row/column cells as <th> (with a scope). — REDACTED:416

What to do

• Declare <html lang>, a document <title> and a <main> landmark, keep headings in order, leave zoom enabled, title iframes and drop meta-refresh.

Other · Accessibility — Whether interactive behaviour is keyboard-reachable — no click handler on a non-interactive element lacking a role, tabindex and key handler, no element the repo's own CSS styles `cursor: pointer` without giving it any of the three, no unfocusable element whose only binding is a mouse enter/leave pair or a double-click, no positive tabindex, no href-less anchor, no placeholder-href (#/javascript) link acting as a button. Static markup readiness, not a WCAG conformance claim.

Method: Static markup-model scan: click handlers on non-interactive elements lacking role+tabindex+key handler, positive tabindex values, and href-less anchors. Components skipped, spreads suppressed. Deterministic, hard fact per element.

Coverage: Population: elements that pose a keyboard-semantics question, in the PARSED MARKUP files only (.html/.htm/.cshtml/.razor/.vue/.svelte/.jsx/.tsx) — a non-natively-interactive element carrying a click handler, a double-click or hover enter/leave binding, a pointer-only gesture on a tabindex="0" element, or `cursor: pointer` from the repo's own CSS; an href-less or placeholder-href (#, javascript:) anchor; and any element with a positive tabindex. Natively interactive elements used correctly (<button>, <a href>, form controls) are NOT in the population — there is nothing to judge — so a page of nothing but correct controls gives AC4 nothing to measure. Keyboard reachability is judged from the markup, never from a rendered page. ★ An interactive element declared in a tagged-template (html`…`) or hyperscript frontend is NOT in this population — no producer reads either — so an empty population is reported as an analyzer gap, never as "this repository has no interactive elements".

• A click handler on a plain element isn't keyboard-operable. Use a <button>, or add role + tabIndex={0} + a key handler. (×2) — REDACTED:159, studio/frontend/src/components/image-viewer.tsx:261
• This <div> binds mousedown and double-click — pointer gestures with no keyboard form, a double-click among them. No click, no key handler, and no tabIndex={0}, so it cannot even take focus. Nothing synthesises a double-click from the keyboard: Enter and Space fire a single click, and only on a native <button> or <a href>. So this behaviour has no keyboard path at all. Expose the same action on a real control, or add role + tabIndex={0} + a key handler alongside the double-click. (×2) — studio/frontend/src/components/tauri/window-titlebar.tsx:519, studio/frontend/src/components/tauri/window-titlebar.tsx:533
• This <div role="alertdialog"> is a modal dialog, but nothing moves focus into it when it opens and nothing keeps focus inside it while it is open, so a keyboard or screen-reader user keeps acting on the content behind it (WCAG 2.4.3). Use a native <dialog> opened with showModal(), a dialog component that traps focus, or move focus to the dialog on open, trap Tab inside it (or make the page behind inert) and return focus to the control that opened it on close. — studio/frontend/src/features/browser/agent/agent-bar.tsx:28
• A click handler on a plain element with no role and no tabindex: even where another element's key handler can reach it, assistive tech can neither focus it nor announce what it is. Use a <button>, or add role + tabIndex={0}. (×5) — studio/frontend/src/features/browser/video-file.tsx:162, studio/frontend/src/features/chat/components/deep-research-composer-button.tsx:117, studio/frontend/src/features/chat/projects-page.tsx:837, …
• This <span> carries its own click handler, but it sits inside a <button> that performs a different action. A keyboard user reaches the <button> and activating it runs the <button>'s handler, never this one, so this action has no keyboard path at all. Role and tabindex cannot supply one: HTML forbids a <button> from containing interactive content or any descendant with a tabindex, so the browser will not put this element in the tab order however it is marked up. Move it out of the <button> and make it a sibling <button>, so each action has its own control. (×2) — studio/frontend/src/features/chat/components/deep-research-composer-button.tsx:297, REDACTED:1479

What to do

• Make custom controls keyboard-operable (role + tabindex + key handler), drop positive tabindex, and give anchors a real href.

Other · Accessibility — Whether ARIA is used correctly — valid non-abstract roles, the ARIA state a role requires, valid (non-misspelled) aria-* attribute names, in-enum values for token-typed aria-* attributes, and no aria-hidden on (or wrapping) a focusable element. Static markup readiness, not a WCAG conformance claim.

Method: Static markup-model scan: role values checked against the WAI-ARIA role set (abstract/invalid flagged), required ARIA state for a role, and aria-hidden on a focusable element. Deterministic, role/attribute level.

Coverage: Population: elements in the PARSED MARKUP files (.html/.htm/.cshtml/.razor/.vue/.svelte/.jsx/.tsx) that carry a role or an aria-* attribute; roles and token values are checked against the ARIA enums exhaustively within that set. An expression-valued (dynamic) role or aria-* value is skipped rather than guessed, and markup built in script — tagged-template (html`…`) UIs and hyperscript DOM factories — is not read at all.

• This is not a valid WAI-ARIA attribute, so assistive technology ignores it (usually a typo). Use the correct aria-* attribute. — studio/frontend/src/features/settings/components/embedding-model-picker.tsx:279

What to do

• Use valid, non-abstract roles, supply the ARIA state each role requires, and never put aria-hidden on a focusable element.

Other · Accessibility — Whether accessibility is ENFORCED in the toolchain — an accessibility checker configured over the markup (an a11y lint rule set, e.g. eslint-plugin-jsx-a11y or vuejs-accessibility where the project lints JavaScript) and an automated accessibility assertion wired into tests or CI (axe/pa11y/Lighthouse or an equivalent) — on the Documented→Verified→Prevented ladder.

Method: Repo config/CI scan: an accessibility checker configured over the markup (an a11y lint rule set such as eslint-plugin-jsx-a11y / vuejs-accessibility where JavaScript is linted) and an automated accessibility assertion in tests or CI (axe/pa11y/Lighthouse or equivalent), graded on the Documented→Verified→Prevented rungs. Deterministic, presence/rung detection.

Coverage: Population: the repository's own tooling configuration — lint config, test and CI files — NOT the markup. It is read for a configured accessibility checker and an automated accessibility assertion (axe/pa11y/Lighthouse, or a native-toolkit equivalent), and it credits an INVOCATION, never a mention: a licence filename, an import comment or a doc reference earns no rung. Enforcement configured entirely outside the repository leaves no evidence here and cannot be credited.

• An a11y linter is configured (Documented) but there's no automated a11y test or CI gate. Add your test runner's axe binding (jest-axe, vitest-axe, cypress-axe or @axe-core/playwright) to verify at runtime, then gate it in CI. What was searched, so you can tell an absence from a miss: the 705 markup file(s) this pass actually assessed, the linter configuration checked in beside them, and this repository's test and CI files — matched by name against the accessibility checkers this dimension carries. An audit run outside the repository, a hosted scanner, or a check whose name is not one of those, is not seen here.

What to do

• Accessibility linting is already configured — assert with your test runner's axe binding (jest-axe, vitest-axe, cypress-axe or @axe-core/playwright) in tests, then gate axe/pa11y/Lighthouse in CI.

Other · Architecture — How the codebase splits by code ROLE — domain, application, infrastructure, test, generated. The significance map behind the knowledge/coupling weighting, and a DDD signal in its own right: a thin domain core under fat infrastructure is the anemic-domain smell, quantified. How each file's role is decided, because the split is only as good as that: a generated name or a build-output tree makes it Generated, a test project makes it Test, and otherwise the file's NAMESPACE and PATH words are matched against fixed vocabularies in a fixed ORDER — domain, then infrastructure, then application — so a file whose words hit two layers is counted under the earlier one. A production file matching none of them counts as application, so that share reads 'application or unclassified' rather than a measured application layer. Roles come from naming convention, never from what the code does. On this repository the split was taken from the source tree on disk rather than from a loaded .NET workspace, so a file's role is decided by its PATH segments alone — no declared namespace was available to add to the evidence — and generated output is excluded from the census entirely rather than counted as a generated share.

Method: Roslyn line-count by code ROLE: every source file classified Domain/Application/Infrastructure/Test/Generated by namespace + path convention (the shared CodeRoleClassifier), then significant lines summed per role. Deterministic; the advisory score is the business-logic (domain+application) share of production code.

Coverage: Population: ALL source files, each bucketed into ONE of five roles (Domain/Application/Infrastructure/Test/Generated) by namespace + path convention — a file whose layer isn't named in the convention falls to Application (the neutral default), and the split is line-count, not semantic depth or business value.

What to do

• The domain core is a small share of production code, but most of the rest matched no layer vocabulary at all — so this is not yet an anemic-domain finding. The namespace/path convention could not place that code, which makes the composition above a statement about the naming, not about the design. Name the layers (or check that the repository's conventions differ from the ones this check knows) before reading a thin domain into it.

Other · Architecture — Whether the project-reference graph is acyclic (cycles block independent build/deploy and signal eroding boundaries).

Method: Project reference cycles via elementary-DFS over real .csproj references, using the engine shared with D5/D7; cyclic versus acyclic. Exhaustive, deterministic.

Other · Architecture — Whether the codebase has a recognisable, scale-appropriate structure (a named architectural style, or modular enough for its size) rather than being an ad-hoc ball of mud.

Method: Roslyn plus csproj analysis: architecture style detection (DDD, clean, vertical-slice, CQRS) and structure fitness for repo size. Deterministic.

Other · Architecture — Whether interfaces stay focused rather than fat — the Interface-Segregation principle (SOLID 'I').

Method: Roslyn scan: public interface declared-member counts (accessors fold into their property/event); fat-interface threshold (over 15 declared members) flagged per type. Each finding also reports the distinct-OPERATION count — members counted by name, so an overload group counts once — which decides whether it states the caller-side ISP harm or the implementer-side burden of an overload set. TypeScript: exported interfaces that declare behaviour, counted over their REQUIRED operations (method signatures and function-typed properties with no `?`, once per signature) — data-shape interfaces and ambient .d.ts declarations are outside the population; JavaScript declares no interfaces and is not applicable. Deterministic, type-level.

• `ChatPreferencesState` declares 19 required operations: `setPlainTextComposer`, `setShowContextWindowUsage`, `setSendShortcut`, `setFollowUpBehavior`, `setConfirmDeleteChats`, `setAlwaysDeleteChatFiles`, `setShowModelDisclaimer`, `setShowResponseModel`, `setShowReplyTimes`, `setShowGenerationDetails`, `setShowInlineEditResponse`, `setThinkingVisibility`, `setToolVisibility`, `setFoldToolActivityIntoThinking`, `setPastedTextMinChars`, `setAutoScrollWhileGenerating`, `setShowScrollToBottomButton`, `setNotifyWhenReplyFinishes`, `setFilesPanelCollapsed`. Counted over its REQUIRED operations — a method signature or a function-typed property with no `?`, once per signature; optional members and data properties are not counted, because no implementer is forced to write them. A wide interface forces every implementer and caller to depend on methods they don't use (the Interface-Segregation 'I' in SOLID). Split it into focused role-interfaces. — studio/frontend/src/features/chat/stores/chat-preferences-store.ts:36
• `ChatsActions` declares 28 required operations: `toggleSelected`, `open`, `rename`, `regenerateTitle`, `togglePin`, `setFavorite`, `toggleFavoriteProject`, `toggleFavoriteSection`, `fork`, `move`, `moveProject`, `viewSection`, `newChatInSection`, `newProjectInSection`, `renameSection`, `removeSection`, `exportSection`, `archive`, `unarchive`, `exportChats`, `remove`, `viewProject`, `filterProject`, `newChatIn`, `editProject`, `togglePinProject`, `exportProject`, `deleteProject`. Counted over its REQUIRED operations — a method signature or a function-typed property with no `?`, once per signature; optional members and data properties are not counted, because no implementer is forced to write them. A wide interface forces every implementer and caller to depend on methods they don't use (the Interface-Segregation 'I' in SOLID). Split it into focused role-interfaces. — studio/frontend/src/features/library/chats/chats-items.tsx:121
• `VoiceSettingsState` declares 26 required operations: `setMicDeviceId`, `setDictationEngine`, `setSttModel`, `setSttGgufVariant`, `setSttDevice`, `setSttProviderId`, `setSttProviderModel`, `setDictationLanguage`, `addDictionaryEntry`, `updateDictionaryEntry`, `commitDictionaryEntry`, `removeDictionaryEntry`, `addRecentDictation`, `removeRecentDictation`, `clearRecentDictations`, `setTtsEnabled`, `setAutoReadAloud`, `setTtsEngine`, `setTtsProviderId`, `setTtsProviderModel`, `setTtsProviderVoice`, `setTtsStudioVoiceId`, `setTtsVoiceURI`, `setTtsRate`, `setTtsPitch`, `setTtsVolume`. Counted over its REQUIRED operations — a method signature or a function-typed property with no `?`, once per signature; optional members and data properties are not counted, because no implementer is forced to write them. A wide interface forces every implementer and caller to depend on methods they don't use (the Interface-Segregation 'I' in SOLID). Split it into focused role-interfaces. — REDACTED:105
• `TrainingConfigActions` declares 80 required operations: `setSelectedModel`, `selectTrainingModel`, `setSelectedModelCacheReference`, `clearSelectedModelCacheReference`, `clearSelectedDatasetCacheReference`, `setSelectedDatasetCacheReference`, `setModelSubfolder`, `setTrainAsDecision`, `setProjectName`, `ensureModelDefaultsLoaded`, `ensureDatasetChecked`, `setTrainingMethod`, `selectHfDataset`, `selectLocalDataset`, `selectS3Source`, `restoreBrowseDatasetSource`, `setDatasetFormat`, `setDataset`, `setDatasetSubset`, `setDatasetSplit`, `setDatasetEvalSplit`, `setDatasetStreaming`, `setManualDatasetOptionsValid`, `markManualDatasetOptionsEdited`, `setDatasetManualMapping`, `setDatasetAdvisorFields`, `setDatasetSliceStart`, `setDatasetSliceEnd`, `setUploadedEvalFile`, `setEpochs`, `setContextLength`, `setVisionImageSize`, `setLearningRate`, `setEmbeddingLearningRate`, `setOptimizerType`, `setLrSchedulerType`, `setLoraRank`, `setLoraAlpha`, `setLoraDropout`, `setLoraVariant`, and 40 more. Counted over its REQUIRED operations — a method signature or a function-typed property with no `?`, once per signature; optional members and data properties are not counted, because no implementer is forced to write them. A wide interface forces every implementer and caller to depend on methods they don't use (the Interface-Segregation 'I' in SOLID). Split it into focused role-interfaces. — studio/frontend/src/features/training/types/config.ts:212
• `TrainingRuntimeActions` declares 17 required operations: `setStopRequested`, `setHydrating`, `setHasHydrated`, `tryBeginStarting`, `setStarting`, `setStartError`, `setStartResources`, `setSseConnected`, `setLastEventId`, `resetRuntime`, `applyStatus`, `applyMetrics`, `applyProgress`, `setStartPending`, `setRuntimeError`, `setSelectedHistoryRunId`, `setCurrentRunViewActive`. Counted over its REQUIRED operations — a method signature or a function-typed property with no `?`, once per signature; optional members and data properties are not counted, because no implementer is forced to write them. A wide interface forces every implementer and caller to depend on methods they don't use (the Interface-Segregation 'I' in SOLID). Split it into focused role-interfaces. — studio/frontend/src/features/training/types/runtime.ts:131

What to do

• Split fat interfaces into focused role-interfaces so clients depend only on what they use.

Other · Architecture — Whether feature slices stay independent (no direct cross-slice references) — the discipline that makes vertical-slice architecture pay off.

Method: Roslyn scan (vertical-slice gated): feature slices resolved from namespaces (.Features.*, .Slices.*) or project names; cross-slice type references detected. Deterministic, traceable.

• `entryDuration` (slice 'api-monitor') depends on `ApiMonitorEntry` from slice 'chat'. (×2) — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:82, studio/frontend/src/features/api-monitor/api-monitor-page.tsx:195
• `SERVER_STATUS` (slice 'api-monitor') depends on `ApiMonitorResponse` from slice 'chat'. (×2) — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:90, studio/frontend/src/features/api-monitor/api-monitor-page.tsx:123
• `ApiMonitorOverlay` (slice 'api-monitor') depends on `useShortcut` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:153
• `ApiMonitorOverlay` (slice 'api-monitor') depends on `getApiMonitor` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:199
• `ApiMonitorPanel` (slice 'api-monitor') depends on `useChatRuntimeStore` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:309
• `ApiMonitorPanel` (slice 'api-monitor') depends on `useMonitorFrameStore` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:352
• `ApiMonitorPanel` (slice 'api-monitor') depends on `FIND_PORTAL_ATTRIBUTE` from slice 'find-in-page'. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:397
• `servedModels` (slice 'api-monitor') depends on `modelIdsMatch` from slice 'hub'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:260
• `openApiSettings` (slice 'api-monitor') depends on `useSettingsDialogStore` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:290
• `ApiMonitorPage` (slice 'api-monitor') depends on `useGuidedTourController` from slice 'tour'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:890
• `ApiMonitorPage` (slice 'api-monitor') depends on `useChatModelRuntime` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:920
• `ApiMonitorPage` (slice 'api-monitor') depends on `useChatRuntimeStore` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:921
• `ApiMonitorPage` (slice 'api-monitor') depends on `ModelConfigHandoffRequest` from slice 'model-picker'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:939
• `ApiMonitorPage` (slice 'api-monitor') depends on `loadLanAccess` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:948
• `ApiMonitorPage` (slice 'api-monitor') depends on `publishLanUrls` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:949
• `ApiMonitorPage` (slice 'api-monitor') depends on `getInferenceStatus` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:963
• `ApiMonitorPage` (slice 'api-monitor') depends on `resolveInferenceCheckpointId` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:964
• `ApiMonitorPage` (slice 'api-monitor') depends on `unloadModel` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:977
• `ApiMonitorPage` (slice 'api-monitor') depends on `isExternalModelId` from slice 'chat'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:986
• `ApiMonitorPage` (slice 'api-monitor') depends on `remoteApiOrigin` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:1147
• `ApiMonitorPage` (slice 'api-monitor') depends on `createModelConfigHandoffRequestId` from slice 'model-picker'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:1197
• `ApiMonitorPage` (slice 'api-monitor') depends on `GuidedTour` from slice 'tour'. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:1303
• `curlExample` (slice 'api-monitor') depends on `psSingle` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-sampling-example.ts:149
• `curlExample` (slice 'api-monitor') depends on `shSingle` from slice 'settings'. — studio/frontend/src/features/api-monitor/api-sampling-example.ts:159
• `ApiExampleChoices` (slice 'api-monitor') depends on `ExampleOs` from slice 'settings'. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:22
• `useApiExampleChoices` (slice 'api-monitor') depends on `useExampleOs` from slice 'settings'. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:33
• `useApiExampleChoices` (slice 'api-monitor') depends on `useSettingsPanelPrefsStore` from slice 'settings'. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:34
• `ApiExampleBox` (slice 'api-monitor') depends on `ExampleCodeBox` from slice 'settings'. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:61
• `ApiExampleBox` (slice 'api-monitor') depends on `ExampleToggle` from slice 'settings'. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:67
• `ApiExampleDialog` (slice 'api-monitor') depends on `authFetch` from slice 'auth'. — studio/frontend/src/features/api-monitor/components/api-example-dialog.tsx:39
• `loadErrorMessage` (slice 'api-monitor') depends on `useChatRuntimeStore` from slice 'chat'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:56
• `startStagedDownload` (slice 'api-monitor') depends on `ModelSelectorChangeMeta` from slice 'model-picker'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:67
• `startStagedDownload` (slice 'api-monitor') depends on `downloadManager` from slice 'hub'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:69
• `startStagedDownload` (slice 'api-monitor') depends on `DOWNLOAD_KIND` from slice 'hub'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:70
• `SelectModelInput` (slice 'api-monitor') depends on `useChatModelRuntime` from slice 'chat'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:93
• `localSelection` (slice 'api-monitor') depends on `resolveResidentInitialConfig` from slice 'model-picker'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:100
• `localSelection` (slice 'api-monitor') depends on `currentRuntimePerModelConfig` from slice 'model-picker'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:124
• `useLoadAfterDownload` (slice 'api-monitor') depends on `useRepoDownload` from slice 'hub'. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:154

What to do

• Keep slices independent: share cross-slice needs via an explicit contract/shared-kernel, not direct references between slices.

Other · Architecture — Whether read (query) handlers stay side-effect-free — a query that writes persistent state or raises events breaks CQS and makes reads unsafe to retry, cache, or route to a read replica.

Method: Roslyn scan: CQRS handlers classified query-vs-command by interface (IQueryHandler/ICommandHandler/IRequestHandler<TQuery,TResult>) and name convention (*Query/Get*/Find* vs *Command); each query handler's body checked for persistent-state writes (SaveChanges/repository Add-Update) or event publishes by resolved invocation. Deterministic, type-level, exhaustive over the detected handlers.

Coverage: Population: CQRS handlers identified by IQueryHandler/ICommandHandler/IRequestHandler interface + *Query/Get*/Find*/*Command NAME convention; query purity then checked exhaustively within that set — a query handler using neither convention is invisible, and mutation is a resolved persistence/publish CALL, not full dataflow.

Other · Security — Whether access is authorized by default — a framework authorization attribute/decorator or policy, or imperative guard methods (throw-on-violation) called from handlers.

Method: Roslyn scan: [Authorize] usage and authorization policies, plus imperative throw-on-violation guard methods detected via syntax. Deterministic.

Other · Event Sourcing — Whether the recovery-replay fold is deterministic — state derived purely from each event's own fields, no clock/random/IO on the replay path.

Method: Roslyn syntax scan (event-sourcing gated): Apply/When folds checked for forbidden tokens (DateTime.Now, Guid.NewGuid, Random, IO), stripped of comments/strings. Deterministic, hard fact per fold.

Other · Event Sourcing — Whether domain events stay immutable (private fields, set once) rather than carrying mutable state.

Method: Roslyn scan (event-sourcing gated): persisted events checked for public setters; immutability verified per property/field. Deterministic, hard fact.

Maturity · Maturity — Whether the repo and its projects have a README, and whether it's substantive and current.

Method: Filesystem scan: README presence, word count, and headings for depth; git history for staleness. Exhaustive across root and project dirs, deterministic.

What to do

• Explain in the README how to run the tests. Add a section called Testing to the README in the repository's top folder, saying how to run the test suite.
• Add an 'Architecture' / 'How it works' section to the root README — the high-level shape.
• Add a README to the 6 of 7 project(s) that lack one — worth up to 1.7 pts.

Maturity · Maturity — Whether key decisions (ADRs) and the high-level shape (C4/diagrams) are written down.

Method: Filesystem scan: ADR folder/naming conventions or content, plus Mermaid/PlantUML/C4/architecture.md discovery. Exhaustive, deterministic.

• No Architecture Decision Records found — no conventional ADR directory, no numbered `NNNN-title` documents in any markup this check reads, and nothing ADR-shaped by content. Design rationale recorded elsewhere (a design-notes tree, a mailing list, pull-request discussion) is not visible to this check and is not re-findable per decision, so a future maintainer cannot ask why one choice was made and get an answer.

What to do

• Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them.

Maturity · Maturity — Whether the repo is organised deliberately — src/test separation and consistent project naming.

Method: Filesystem scan: src/test folder separation and namespace-prefix consistency (majority RootNamespace agreement). Exhaustive across projects, deterministic.

• Production code isn't grouped under a src/ folder — it's spread across several top-level directories, so there's no one place that says 'this is the product'.

What to do

• Group production code under src/ (or split deliberately, e.g. backend/ + frontend/) so production and tooling code aren't mixed at the root.

Maturity · Maturity — Whether the README actually describes the code that exists (LLM-judged, advisory).

Method: Judged by language model at low temperature: README accuracy versus actual projects, within a disclosed tolerance. Advisory, not a measured number.

• README.md names `unsloth/unsloth`, but no file or folder `unsloth/unsloth` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — README.md:73
• README.md names `unsloth/unsloth-rocm`, but no file or folder `unsloth/unsloth-rocm` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — README.md:200
• unsloth/kernels/moe/README.md names `grouped_gemm/tests/test_grouped_gemm.py`, but no file or folder `grouped_gemm/tests/test_grouped_gemm.py` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — unsloth/kernels/moe/README.md:42
• unsloth/kernels/moe/README.md names `grouped_gemm/tests/test_qwen3_moe.py`, but no file or folder `grouped_gemm/tests/test_qwen3_moe.py` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — unsloth/kernels/moe/README.md:44
• unsloth/kernels/moe/README.md names `grouped_gemm/benchmark/benchmark_fused_moe.py`, but no file or folder `grouped_gemm/benchmark/benchmark_fused_moe.py` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — unsloth/kernels/moe/README.md:47
• tests/kaggle/t4_smoke/references/README.md names `unsloth/Qwen2.5-0.5B-Instruct-unsloth-bnb-4bit`, but no file or folder `unsloth/Qwen2.5-0.5B-Instruct-unsloth-bnb-4bit` exists. Checked without a model against the repository: its files and folders, the scripts and dependencies of every package.json, the projects, namespaces, package references and target frameworks of the .NET build, the SDK and Node pins, and the identifiers of every file that is not a document; the documents' own text never counts. A footprint generated at build time, read only by a platform, or living outside this repository is not seen, and this row is then wrong. — tests/kaggle/t4_smoke/references/README.md:247

What to do

• Reconcile the README with reality: README.md names `unsloth/unsloth`, but no file or folder `unsloth/unsloth` exists; README.md names `unsloth/unsloth-rocm`, but no file or folder `unsloth/unsloth-rocm` exists; unsloth/kernels/moe/README.md names `grouped_gemm/tests/test_grouped_gemm.py`, but no file or folder `grouped_gemm/tests/test_grouped_gemm.py` exists; unsloth/kernels/moe/README.md names `grouped_gemm/tests/test_qwen3_moe.py`, but no file or folder `grouped_gemm/tests/test_qwen3_moe.py` exists; unsloth/kernels/moe/README.md names `grouped_gemm/benchmark/benchmark_fused_moe.py`, but no file or folder `grouped_gemm/benchmark/benchmark_fused_moe.py` exists.

Readiness · Readiness — Whether an automated pipeline builds and tests every change.

Method: Filesystem scan: CI workflow files (.github/workflows, .gitlab-ci.yml, etc.) for build and test stages. Exhaustive, deterministic.

Readiness · Readiness — For a library: a deliberate (small) public API surface and explicit semantic versioning so consumers can depend on it safely.

Method: Roslyn scan: public API surface area and semantic-versioning markers (SemVer attributes, changelog entries) for libraries; off .NET, a library is the ecosystem's publication act (an npm package that is not private and names an entry point, a PyPI distribution with a build system, a Rust library crate, a Maven/Gradle module that publishes, a Go module with no package main, a gemspec, a Composer library, a SwiftPM library product, a pub.dev or Hex package), its surface is the share of types the language model records as public (Rust, Swift, Java, Kotlin, Go, Dart; not measured where the model records no type visibility or, as in TypeScript, only module-level export), and its version is read from the manifest, a semver CHANGELOG, release tooling or semver git tags. Exhaustive, deterministic.

Readiness · Readiness — Whether the code is diagnosable in production — structured logging, tracing/metrics, health checks.

Method: Filesystem/Roslyn scan: structured-logging frameworks (Serilog, NLog), OpenTelemetry, and health-check endpoint patterns. Exhaustive, deterministic.

What to do

• Add OpenTelemetry tracing/metrics (opentelemetry-sdk) so requests are traceable across the system, not just health-probable.

Readiness · Readiness — Whether SAST, secret/dependency scanning and performance benchmarking are wired in (presence, not runtime).

Method: Filesystem scan: SAST configuration, dependency-update automation, secret scanning, and a benchmark harness or benchmark step — in this repository's own ecosystem. Exhaustive, deterministic.

Readiness · Readiness — Whether releases are automated and safely reversible (probes, rolling updates, approval gates) — from manifests/pipeline files, not the live environment.

Method: Filesystem scan: deployment manifests/IaC (K8s YAML, Helm, Terraform) for rolling updates, probes, approval gates, migration hooks. Exhaustive, deterministic.

Readiness · Readiness — Whether releases are traceable — a maintained changelog and explicit version stamping.

Method: Filesystem scan: changelog file presence and version tags in csproj or git tags. Exhaustive, deterministic.

• No CHANGELOG/HISTORY/RELEASES file — what shipped when isn't easy to reconstruct for support or audit. (Versioning/tagging makes releases traceable, but a changelog records the what.)

What to do

• Keep a changelog (e.g. Keep-a-Changelog) recording what shipped in each release.

Readiness · Readiness — Whether outbound HTTP calls are wrapped in resilience (retry/timeout/circuit-breaker) so a failing dependency doesn't cascade.

Method: Source scan: outbound HTTP clients and what bounds them — resilience handlers (Polly, AddStandardResilienceHandler) on .NET; on Go, the JVM, Python, JavaScript/TypeScript, Ruby, PHP, Rust, Elixir, Swift, Dart and Erlang, a timeout, deadline, retry or breaker beside each call, or a process-wide client default (a framework-wide deadline such as Drupal core's, Laravel's or actix's awc counts). Exhaustive, deterministic.

• `httpx.AsyncClient(` makes an outbound HTTP call, and nothing bounds it: no timeout, deadline, retry or circuit breaker is set for it here, and the client has no process-wide default. A slow or failing dependency will hold this service's request, thread or connection until the call gives up on its own — or never, for a client with no default timeout. (×2) — studio/backend/core/inference/inference_gate.py:77, studio/backend/core/inference/llama_http.py:25

What to do

• Pass `timeout=` on every `requests`/`httpx` call (requests has no default and will wait forever), mount an `HTTPAdapter(max_retries=Retry(...))` on shared sessions, and wrap calls to flaky dependencies in `tenacity` retries or a `pybreaker` circuit breaker.

Readiness · Performance — Whether asynchronous code stays responsive — it avoids sync-over-async blocking (a .NET .Wait()/.GetAwaiter().GetResult(), a time.sleep or blocking HTTP call inside a Python coroutine, a *Sync call inside an async JavaScript function, block_on inside a Rust async fn, runBlocking inside a Kotlin suspend function, block() inside a Reactor publisher) that stalls a thread or event loop and risks deadlock, and, where the code is a reusable library on .NET, awaits with ConfigureAwait(false) so it never captures and stalls its caller's context.

Method: Production-source scan: sync-over-async blocking counted everywhere — .Wait()/.GetAwaiter().GetResult() in .NET; off .NET, read from the language model, a blocking call inside an async function (Python, TS/JS, Rust, Kotlin, Swift, Dart) or inside a Java method returning a Reactor Mono/Flux or a Scala method returning a Future or effect — and, for a .NET library with ≥5 awaits, the share of awaits using ConfigureAwait(false). Deterministic, syntax/text detection.

• `unzipSync` in `readOpenDocumentXmlFiles` blocks the event loop until it finishes — a blocking call inside async code stalls the event loop — every other request waits for as long as it runs. — studio/frontend/src/features/chat/open-document.ts:96
• `readFileSync` in `bundleOne` blocks the event loop until it finishes — a blocking call inside async code stalls the event loop — every other request waits for as long as it runs. — studio/frontend/vite-preview-runtime.ts:190

What to do

• TypeScript/JavaScript: use the promise APIs (fs/promises, a promisified child_process.execFile, the async zlib/crypto functions) and await them instead of the *Sync variants.

React / JS · Code Health — How much of the frontend is typed TypeScript vs untyped JavaScript.

Method: Frontend file inventory: the share of typed TypeScript vs untyped JavaScript across the source tree. Deterministic, exhaustive over frontend files.

What to do

• Migrate the remaining .js/.jsx files to TypeScript.

React / JS · Code Health — Near-exact copy-pasted blocks of substantial extent across the frontend (the D4 clone algorithm over JS/TS tokens, D-386): a block is reported only where its copies still agree on most of their own identifiers and literals, or were renamed as they were pasted but kept most of their constants, and where the copies carry enough code to stand on their own or the copied extent reaches 30 lines — so a re-implementation sharing neither names nor values, and a small pasted declaration, are both found and deliberately not reported, and a clean R10 is not a claim that nothing was copied.

Method: Near-exact copy-pasted blocks of substantial extent across the frontend (the D4 clone algorithm run over JS/TS tokens). Masking finds the candidates; a block is reported when its copies still agree on most of their own identifiers and literals, or when a renamed copy still agrees on most of its constants, AND the copies carry enough code to stand on their own — or when the copied extent reaches 30 lines. So a re-implementation sharing neither names nor values, and a small pasted declaration, are deliberately not counted. Deterministic.

• 455 duplicated blocks under studio/frontend/src/features/ have copies in at least two of the sibling directories api-monitor, audio, auth, browser, chat, data-recipes (+24 more sibling(s) not listed) — 188 of them are reported below, and 267 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 455 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 455 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 455 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/dataset-picker/components/dataset-selector-lists.tsx:58
• 165 duplicated blocks under studio/frontend/src/ have copies in at least two of the sibling directories app, components, features, hooks, lib — 85 of them are reported below, and 80 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 165 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 165 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 165 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/components/assistant-ui/thread.tsx:7421
• 42 duplicated blocks under studio/frontend/src/features/settings/ have copies in at least two of the sibling directories components, hooks, lib, tabs — 22 of them are reported below, and 20 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 42 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 42 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 42 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/settings/components/usage-examples.tsx:953
• 17 duplicated blocks under studio/frontend/src/features/recipe-studio/ have copies in at least two of the sibling directories components, dialogs, easy, executions, hooks, stores (+1 more sibling(s) not listed) — 9 of them are reported below, and 8 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 17 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 17 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 17 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/recipe-studio/stores/helpers/model-infra-layout.ts:29
• 16 duplicated blocks under studio/frontend/src/features/recipe-studio/dialogs/ have copies in at least two of the sibling directories expression, llm, markdown-note, models, samplers, seed (+2 more sibling(s) not listed) — 7 of them are reported below, and 9 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 16 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 16 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 16 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/recipe-studio/dialogs/markdown-note/markdown-note-dialog.tsx:11
• 11 duplicated blocks under studio/frontend/src/features/chat/ have copies in at least two of the sibling directories api, artifacts, hooks, lib, prompt-storage, stores (+2 more sibling(s) not listed) — 6 of them are reported below, and 5 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 11 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 11 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 11 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/chat/hooks/use-chat-model-runtime.ts:2987
• 5 duplicated blocks under studio/frontend/src/features/audio/ have copies in at least two of the sibling directories components, hooks, pages, tools — 2 of them are reported below, and 3 are counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 5 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 5 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 5 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/audio/components/save-voice-dialog.tsx:133
• 5 duplicated blocks under studio/frontend/src/features/hub/ have copies in at least two of the sibling directories download-manager, hooks, inventory, lib, stores — 4 of them are reported below, and 1 is counted here but not reported individually: those copies match on shape but no longer clear R10's bar for an individually reported row — either they kept neither their own names nor their values, or what was copied is too small to stand on its own (it reports near-exact duplication only, and only of substantial extent). What this row states is the concentration, which the detector measured over all 5 and which does not depend on how exactly each block's copies still match. That concentration is one structural fact, not 5 local ones: the siblings replicate behaviour none of them owns, which is the shape of a missing shared module — a common library every sibling imports — rather than 5 separate extractions. Check first whether the siblings are deliberately standalone deliverables (scaffold templates, demo apps that must stay copy-pasteable); where they are, the duplication is the design and the per-block rows are the ones to act on. — studio/frontend/src/features/hub/hooks/use-hub-model-search.ts:200
• studio/frontend/src/features/audio/hooks/use-convert-generation.ts:89 · studio/frontend/src/features/audio/hooks/use-edit-generation.ts:94 — the two spans are one implementation copied and then locally edited — 1069 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/audio/hooks/use-convert-generation.ts:89
• studio/frontend/src/features/settings/components/lan-access-section.tsx:109 · studio/frontend/src/features/settings/components/remote-access-section.tsx:90 — the two spans are one implementation copied and then locally edited — 907 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/settings/components/lan-access-section.tsx:109
• studio/frontend/src/features/dataset-picker/components/dataset-selector-lists.tsx:58 · studio/frontend/src/features/train-model-picker/components/train-model-picker-lists.tsx:69 — the two spans are one implementation copied and then locally edited — 767 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/dataset-picker/components/dataset-selector-lists.tsx:58
• studio/frontend/src/features/library/chats/chats-toolbar.tsx:128 · studio/frontend/src/features/library/components/library-toolbar.tsx:132 — the two spans are one implementation copied and then locally edited — 892 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/library/chats/chats-toolbar.tsx:128
• studio/frontend/src/features/training/lib/resume-training-run.ts:98 · studio/frontend/src/features/training/lib/start-fresh-training-run.ts:116 — the two spans are one implementation copied and then locally edited — 734 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/training/lib/resume-training-run.ts:98
• studio/frontend/src/features/images/images-page.tsx:2709 · studio/frontend/src/features/video/video-page.tsx:2041 — the 2 copies are spread across 2 directories, so the shared home is a decision rather than an obvious spot: check first whether one of them already owns this behaviour, and otherwise put the extracted module somewhere all of the sites already reach rather than making one of them depend on another. — studio/frontend/src/features/images/images-page.tsx:2709
• studio/frontend/src/features/images/images-page.tsx:1065 · studio/frontend/src/features/images/localized-edit-canvas.tsx:35 — the two spans are one implementation copied and then locally edited — 1060 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/images/images-page.tsx:1065
• studio/frontend/src/components/ui/dropdown-menu.tsx:89 · studio/frontend/src/components/ui/menubar.tsx:99 — the two spans are one implementation copied and then locally edited — 515 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/components/ui/dropdown-menu.tsx:89
• studio/frontend/src/features/settings/components/audio-api-examples.tsx:76 · studio/frontend/src/features/settings/components/media-api-examples.tsx:42 — the two spans are one implementation copied and then locally edited — 757 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/settings/components/audio-api-examples.tsx:76
• studio/frontend/src/components/assistant-ui/tool-fallback.tsx:95 · studio/frontend/src/components/assistant-ui/tool-group.tsx:76 — the two spans are one implementation copied and then locally edited — 440 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/components/assistant-ui/tool-fallback.tsx:95
• studio/frontend/src/features/audio/tools/clone-panels.tsx:37 · studio/frontend/src/features/audio/tools/convert-panels.tsx:38 — the two spans are one implementation copied and then locally edited — 528 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/audio/tools/clone-panels.tsx:37
• studio/frontend/src/features/images/images-page.tsx:3475 · studio/frontend/src/features/video/video-page.tsx:2456 — the two spans are one implementation copied and then locally edited — 682 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/images/images-page.tsx:3475
• studio/frontend/src/components/ui/context-menu.tsx:33 · studio/frontend/src/components/ui/dropdown-menu.tsx:19 — the two spans are one implementation copied and then locally edited — 370 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/components/ui/context-menu.tsx:33
• studio/frontend/src/components/ui/dialog.tsx:52 · studio/frontend/src/components/ui/sheet.tsx:59 — the two spans are one implementation copied and then locally edited — 436 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/components/ui/dialog.tsx:52
• studio/frontend/src/features/images/mask-from-text-panel.tsx:69 · studio/frontend/src/features/images/prompt-enhance-panel.tsx:83 — the two spans are one implementation copied and then locally edited — 489 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/images/mask-from-text-panel.tsx:69
• studio/frontend/src/components/assistant-ui/thread.tsx:3431 · studio/frontend/src/components/assistant-ui/thread.tsx:3547 · studio/frontend/src/components/assistant-ui/thread.tsx:3797 — all 3 copies are in the same file, so extract the block into one function there and call it from every one of those sites — resolving only two of them leaves the rest to drift apart the first time one is edited. — studio/frontend/src/components/assistant-ui/thread.tsx:3431
• studio/frontend/src/features/settings/components/sidebar-menu-customizer.tsx:39 · studio/frontend/src/features/settings/components/sidebar-nav-customizer.tsx:48 — the two spans are one implementation copied and then locally edited — 506 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/settings/components/sidebar-menu-customizer.tsx:39
• studio/frontend/src/features/hub/catalog/model-card.tsx:238 · studio/frontend/src/features/hub/catalog/models-catalog-rows.tsx:455 — the two spans are one implementation copied and then locally edited — 390 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/hub/catalog/model-card.tsx:238
• studio/frontend/smoke-heavy-thread-main.tsx:746 · studio/frontend/smoke-thread-weight-main.tsx:302 — the two spans are one implementation copied and then locally edited — 547 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/smoke-heavy-thread-main.tsx:746
• studio/frontend/src/components/ui/context-menu.tsx:202 · studio/frontend/src/components/ui/menubar.tsx:136 — the 2 copies are spread across 2 files, and the CITED SPAN is not a self-contained block — it runs from inside one construct into the next (the tail of a branch plus the head of the following one, a run of switch arms, the end of a declaration plus the list that follows it) rather than covering a whole unit. So do not lift these lines literally: no call can be substituted for a half-open construct. Extract the enclosing repeated UNIT instead — the whole function, component or branch these lines sit in — and where the repetition IS the construct (a run of switch arms, a stack of near-identical declarations) replace it with one table or registry looked up by key rather than a helper each arm calls. The copies still drift apart the first time only one of them is edited, which is why this is reported. — studio/frontend/src/components/ui/context-menu.tsx:202
• studio/frontend/src/features/studio/sections/charts/grad-norm-chart-card.tsx:4 · studio/frontend/src/features/studio/sections/charts/learning-rate-chart-card.tsx:4 — these 2 files are line-for-line copies of one another — 73 lines are identical, in the same order, in every one of them — so this is one fact about the file set, not a block to extract. An edit made to one file and not the others changes behaviour silently, which is the failure a wholesale copy guarantees. Pick one file as the single source and derive the others from it (re-export it, spread it into the local overrides each variant genuinely needs, or generate the copies at build time) — the few lines that differ between the files are exactly the part each variant should still own. Check first whether the copies are deliberately standalone deliverables (a translation file seeded from its sibling and waiting to be translated); where they are, the duplication is the design, and the honest move is to mark the seeded file as untranslated rather than to let it pass as done. — studio/frontend/src/features/studio/sections/charts/grad-norm-chart-card.tsx:4
• studio/frontend/src/features/audio/hooks/use-edit-generation.ts:335 · studio/frontend/src/features/audio/hooks/use-separate-generation.ts:242 — the 2 copies are spread across 2 files, and the CITED SPAN is not a self-contained block — it runs from inside one construct into the next (the tail of a branch plus the head of the following one, a run of switch arms, the end of a declaration plus the list that follows it) rather than covering a whole unit. So do not lift these lines literally: no call can be substituted for a half-open construct. Extract the enclosing repeated UNIT instead — the whole function, component or branch these lines sit in — and where the repetition IS the construct (a run of switch arms, a stack of near-identical declarations) replace it with one table or registry looked up by key rather than a helper each arm calls. The copies still drift apart the first time only one of them is edited, which is why this is reported. — studio/frontend/src/features/audio/hooks/use-edit-generation.ts:335
• studio/frontend/src/features/recipe-studio/dialogs/samplers/datetime-dialog.tsx:36 · studio/frontend/src/features/recipe-studio/dialogs/samplers/timedelta-dialog.tsx:32 — the two spans are one implementation copied and then locally edited — 369 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/recipe-studio/dialogs/samplers/datetime-dialog.tsx:36
• studio/frontend/src/features/recipe-studio/dialogs/samplers/gaussian-dialog.tsx:22 · studio/frontend/src/features/recipe-studio/dialogs/samplers/uniform-dialog.tsx:22 — the 2 copies sit in sibling files in one directory, so check first whether one of them (or an existing module there) already owns this behaviour and the others should call it; otherwise extract it into one module in that directory and have each site call it. — studio/frontend/src/features/recipe-studio/dialogs/samplers/gaussian-dialog.tsx:22
• studio/frontend/src/components/assistant-ui/thread.tsx:7421 · studio/frontend/src/features/chat/shared-composer.tsx:2985 — the 2 copies are spread across 2 files, and the SHAPE of this repetition could not be determined. It is not a run of declarations, a listing, a declaration header, a type body or a slice through a construct — and it was not measured as a run of executable statements either, so this row cannot tell you whether a function can stand where these lines are. Read the two spans before acting, because the move is opposite in the two cases. Where they are statements, the ordinary answer holds: give the shared part one home and call it from each site. Where they turn out to be declarations, a literal's entries, or the cases of an enumeration, there is no call site to call anything from, and collapsing them would delete what each copy pins — a shared base type, a generated set, or one exported constant each site refers to is the move instead, and sometimes the honest answer is that there is nothing to extract at all. Reported because the copies drift apart the first time only one of them is edited, which is true whichever of those they are. — studio/frontend/src/components/assistant-ui/thread.tsx:7421
• studio/frontend/src/features/hub/catalog/dataset-download-section.tsx:49 · studio/frontend/src/features/hub/catalog/safetensors-download-card.tsx:103 — the two spans are one implementation copied and then locally edited — 309 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/hub/catalog/dataset-download-section.tsx:49
• studio/frontend/src/features/library/chats/chats-items.tsx:1482 · studio/frontend/src/features/library/chats/chats-items.tsx:1720 — the two spans are one implementation copied and then locally edited — 338 tokens are still identical, in the same order in both spans, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/library/chats/chats-items.tsx:1482
• REDACTED:8202 · REDACTED:8349 · REDACTED:8483 — all 3 copies are in the same file, and the CITED SPAN is not a self-contained block — it runs from inside one construct into the next (the tail of a branch plus the head of the following one, a run of switch arms, the end of a declaration plus the list that follows it) rather than covering a whole unit. So do not lift these lines literally: no call can be substituted for a half-open construct. Extract the enclosing repeated UNIT instead — the whole function, component or branch these lines sit in — and where the repetition IS the construct (a run of switch arms, a stack of near-identical declarations) replace it with one table or registry looked up by key rather than a helper each arm calls. The copies still drift apart the first time only one of them is edited, which is why this is reported. — REDACTED:8202
• studio/frontend/src/features/hub/catalog/models-table.tsx:357 · studio/frontend/src/features/hub/catalog/models-table.tsx:513 — the two spans are one implementation copied and then locally edited — 289 tokens are still identical, in the same order in both spans, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/hub/catalog/models-table.tsx:357
• studio/frontend/src/components/assistant-ui/thread.tsx:6718 · studio/frontend/src/features/chat/shared-composer.tsx:3101 — the two spans are one implementation copied and then locally edited — 282 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/components/assistant-ui/thread.tsx:6718
• studio/frontend/src/features/browser/downloads-button.tsx:53 · studio/frontend/src/features/browser/internal-pages.tsx:647 — the two spans are one implementation copied and then locally edited — 376 tokens are still identical, in the same order in both files, with only local edits between them. The copies have already begun to drift, which is this row's finding: an edit made to one and not the other changes behaviour silently. Diff the two spans first to learn what genuinely differs, then extract the shared core into one module both sites use, passing the differences in as parameters — or, if one copy exists only because the other could not be imported from its context, make one of them the single source the other is generated or re-exported from. If one copy is no longer reachable, delete it rather than letting it shadow the live one. — studio/frontend/src/features/browser/downloads-button.tsx:53
• studio/frontend/src/features/images/images-page.tsx:893 · studio/frontend/src/features/video/video-page.tsx:562 — the 2 copies are spread across 2 directories, so the shared home is a decision rather than an obvious spot: check first whether one of them already owns this behaviour, and otherwise put the extracted module somewhere all of the sites already reach rather than making one of them depend on another. — studio/frontend/src/features/images/images-page.tsx:893

What to do

• Act on each finding's own remediation rather than one rule: the move depends on what recurs. Where the copies are executable blocks, give the shared part one home and call it from each site; where they are declarations, a listing, a specialisation already delegating to its base, or one shape repeated per entity, there is no call site and the move is a shared type, a generated set or a factory — sometimes there is nothing to extract.

React / JS · Architecture — Conformance to the detected frontend architecture layout (feature-sliced / layered src) plus cross-package deep-import rules (D-386).

Method: Conformance to the detected frontend layout (feature-sliced / layered src) plus cross-package deep-import rules, over the module graph. Deterministic.

• studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:7 imports studio/frontend/src/features/chat/api/chat-api.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:7
• studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:8 imports studio/frontend/src/features/chat/stores/chat-runtime-store.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:8
• studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:13 imports studio/frontend/src/features/find-in-page/lib/find-attributes.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:13
• studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:14 imports studio/frontend/src/features/settings/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-overlay.tsx:14
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:5 imports studio/frontend/src/features/tour/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:5
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:22 imports studio/frontend/src/features/chat/api/chat-api.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:22
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:23 imports studio/frontend/src/features/chat/lib/apply-inference-status-to-store.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:23
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:24 imports studio/frontend/src/features/chat/stores/chat-runtime-store.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:24
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:25 imports studio/frontend/src/features/chat/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:25
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:27 imports studio/frontend/src/features/chat/external-providers.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:27
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:28 imports studio/frontend/src/features/hub/lib/model-identity.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:28
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:29 imports studio/frontend/src/features/model-picker/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:29
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:33 imports studio/frontend/src/features/settings/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:33
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx:38 imports studio/frontend/src/features/settings/api/remote-access-state.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:38
• studio/frontend/src/features/api-monitor/api-sampling-example.ts:4 imports studio/frontend/src/features/settings/components/agent-command.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/api-sampling-example.ts:4
• studio/frontend/src/features/api-monitor/components/api-example-box.tsx:4 imports studio/frontend/src/features/settings/components/api-example-parts.tsx against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:4
• studio/frontend/src/features/api-monitor/components/api-example-box.tsx:10 imports studio/frontend/src/features/settings/stores/settings-panel-prefs-store.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-example-box.tsx:10
• studio/frontend/src/features/api-monitor/components/api-example-dialog.tsx:11 imports studio/frontend/src/features/auth/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-example-dialog.tsx:11
• studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:8 imports studio/frontend/src/features/chat/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:8
• studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:16 imports studio/frontend/src/features/hub/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:16
• studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:23 imports studio/frontend/src/features/model-picker/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:23
• studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:33 imports studio/frontend/src/features/npu/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/api-model-load-controls.tsx:33
• studio/frontend/src/features/api-monitor/components/saved-model-settings.tsx:10 imports studio/frontend/src/features/model-picker/api/model-overrides.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/saved-model-settings.tsx:10
• studio/frontend/src/features/api-monitor/components/saved-model-settings.tsx:16 imports studio/frontend/src/features/model-picker/model-config/per-model-config.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/components/saved-model-settings.tsx:16
• studio/frontend/src/features/api-monitor/forget-model-override.ts:4 imports studio/frontend/src/features/model-picker/api/model-overrides.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/forget-model-override.ts:4
• studio/frontend/src/features/api-monitor/forget-model-override.ts:5 imports studio/frontend/src/features/model-picker/model-config/model-identity.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/forget-model-override.ts:5
• studio/frontend/src/features/api-monitor/slots-settings-request.ts:4 imports studio/frontend/src/features/hub/lib/model-identity.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/slots-settings-request.ts:4
• studio/frontend/src/features/api-monitor/slots-settings-request.ts:5 imports studio/frontend/src/features/model-picker/model-config/model-identity.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/slots-settings-request.ts:5
• studio/frontend/src/features/api-monitor/slots-settings-request.ts:7 imports studio/frontend/src/features/model-picker/model-config/model-config-handoff.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/slots-settings-request.ts:7
• studio/frontend/src/features/api-monitor/use-api-monitor.ts:4 imports studio/frontend/src/features/chat/api/chat-api.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/use-api-monitor.ts:4
• studio/frontend/src/features/api-monitor/use-resident-run-target.ts:4 imports studio/frontend/src/features/chat/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/use-resident-run-target.ts:4
• studio/frontend/src/features/api-monitor/use-resident-run-target.ts:5 imports studio/frontend/src/features/hub/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/use-resident-run-target.ts:5
• studio/frontend/src/features/api-monitor/use-resident-run-target.ts:6 imports studio/frontend/src/features/model-picker/index.ts against the feature-sliced rules. — studio/frontend/src/features/api-monitor/use-resident-run-target.ts:6
• studio/frontend/src/features/audio/api.ts:4 imports studio/frontend/src/features/auth/index.ts against the feature-sliced rules. — studio/frontend/src/features/audio/api.ts:4
• studio/frontend/src/features/audio/audio-options-fields.tsx:13 imports studio/frontend/src/features/chat/index.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-options-fields.tsx:13
• studio/frontend/src/features/audio/audio-page-policy.ts:6 imports studio/frontend/src/features/model-picker/components/model-selector/audio-picker-policy.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-page-policy.ts:6
• studio/frontend/src/features/audio/audio-page.tsx:14 imports studio/frontend/src/features/tour/index.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-page.tsx:14
• studio/frontend/src/features/audio/audio-page.tsx:18 imports studio/frontend/src/features/chat/index.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-page.tsx:18
• studio/frontend/src/features/audio/audio-page.tsx:26 imports studio/frontend/src/features/hub/index.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-page.tsx:26
• studio/frontend/src/features/audio/audio-page.tsx:27 imports studio/frontend/src/features/hub/stores/inventory-events.ts against the feature-sliced rules. — studio/frontend/src/features/audio/audio-page.tsx:27

What to do

• Fix the listed violations: respect the layer/slice order — import only through a slice's public index and never upward into a higher layer (918 rows); import through the target package's declared public entry rather than a deep path into its internals or a relative path into its directory (1 row).

React / JS · Code Health — Per-function cyclomatic/cognitive complexity from the token-level function scanner (D-386) — real branching, not a regex heuristic.

Method: Per-function cyclomatic/cognitive complexity from a token-level function scanner (real branching, not a regex heuristic), computed over every frontend function. Deterministic.

• run has cyclomatic complexity 678 and cognitive complexity 1460; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/api/chat-adapter.ts:4614
• performLoad has cyclomatic complexity 409 and cognitive complexity 453; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/hooks/use-chat-model-runtime.ts:2109
• ImagesPage has cyclomatic complexity 284 and cognitive complexity 253; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/images/images-page.tsx:1650
• (anonymous) has cyclomatic complexity 245 and cognitive complexity 320; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/hooks/use-chat-model-runtime.ts:1295
• applyActiveModelStatusToStore has cyclomatic complexity 241 and cognitive complexity 226; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/lib/apply-inference-status-to-store.ts:172
• ModelConfigPage has cyclomatic complexity 223 and cognitive complexity 184; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/model-picker/components/model-config-page.tsx:2539
• AudioPage has cyclomatic complexity 210 and cognitive complexity 196; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/audio/audio-page.tsx:264
• (anonymous) has cyclomatic complexity 171 and cognitive complexity 261; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/tests/chat-barrel-module-scope-reads.test.ts:556
• ExportPage has cyclomatic complexity 171 and cognitive complexity 139; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/export/export-page.tsx:195
• perform has cyclomatic complexity 170 and cognitive complexity 227; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/browser/agent/executor.ts:732
• DiffusionTrainPanel has cyclomatic complexity 167 and cognitive complexity 139; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/images/train/diffusion-train-panel.tsx:377
• AppSidebar has cyclomatic complexity 163 and cognitive complexity 146; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/components/app-sidebar.tsx:1340
• getConfigErrors has cyclomatic complexity 162 and cognitive complexity 321; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/recipe-studio/utils/validation.ts:14
• ensureModelLoaded has cyclomatic complexity 159 and cognitive complexity 175; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/shared-composer.tsx:1660
• SharedComposer has cyclomatic complexity 151 and cognitive complexity 133; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/shared-composer.tsx:621
• ChatPage has cyclomatic complexity 151 and cognitive complexity 126; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/chat-page.tsx:2661
• GgufDownloadCard has cyclomatic complexity 132 and cognitive complexity 101; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/hub/catalog/gguf-download-card.tsx:657
• (anonymous) has cyclomatic complexity 127 and cognitive complexity 199; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/chat/api/chat-adapter.ts:6761
• VideoGenerator has cyclomatic complexity 127 and cognitive complexity 112; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive sits below cyclomatic here, so much of the count is breadth — arms side by side rather than stacked — and splitting per arm would leave a function per arm; group the work between the checks into named steps instead. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/video/video-page.tsx:1053
• (anonymous) has cyclomatic complexity 125 and cognitive complexity 141; this row is raised above a cyclomatic bar of 10. The two numbers answer different questions and the gap between them is what decides whether to act: cyclomatic counts the independent arms through the body, cognitive counts what it costs to hold them in your head, so nesting and mixed boolean chains raise it while a flat run of independent arms does not. Cognitive is at or above cyclomatic here, so the branching is nested or entangled rather than laid out side by side — extracting each decision into its own named function is the change that pays. Measured by this repository's own parse of the file, so a body assembled at runtime, or generated, is counted as written rather than as it executes. — studio/frontend/src/features/images/images-page.tsx:5341

What to do

• Break down the listed branch-heavy functions; aim P95 cyclomatic ≤ 5.

React / JS · Code Health — How many source files exceed the large-file threshold.

Method: Components/modules exceeding the large-file threshold, counted exhaustively across the frontend source tree. Deterministic.

• studio/frontend/src/components/assistant-ui/thread.tsx has 9153 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/components/assistant-ui/thread.tsx:1
• studio/frontend/src/features/chat/api/chat-adapter.ts has 9109 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/api/chat-adapter.ts:1
• REDACTED has 9062 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — REDACTED:1
• studio/frontend/src/features/images/images-page.tsx has 7546 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/images/images-page.tsx:1
• studio/frontend/src/components/app-sidebar.tsx has 7211 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/components/app-sidebar.tsx:1
• studio/frontend/src/features/chat/stores/chat-runtime-store.ts has 6372 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/stores/chat-runtime-store.ts:1
• studio/frontend/src/features/chat/chat-page.tsx has 4951 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/chat-page.tsx:1
• studio/frontend/src/features/video/video-page.tsx has 4843 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/video/video-page.tsx:1
• studio/frontend/src/features/chat/runtime-provider.tsx has 4467 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/runtime-provider.tsx:1
• studio/frontend/src/features/model-picker/components/model-config-page.tsx has 4234 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/model-picker/components/model-config-page.tsx:1
• studio/frontend/src/features/chat/hooks/use-chat-model-runtime.ts has 4093 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/hooks/use-chat-model-runtime.ts:1
• studio/frontend/src/features/chat/shared-composer.tsx has 3519 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/shared-composer.tsx:1
• studio/frontend/src/features/chat/prompt-storage/prompt-storage-dialog.tsx has 2683 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/prompt-storage/prompt-storage-dialog.tsx:1
• studio/frontend/src/features/chat/chat-providers-dialog.tsx has 2521 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/chat-providers-dialog.tsx:1
• studio/frontend/src/features/images/train/diffusion-train-panel.tsx has 2476 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/images/train/diffusion-train-panel.tsx:1
• studio/frontend/src/features/browser/browser-panel.tsx has 2470 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/browser/browser-panel.tsx:1
• studio/frontend/src/features/audio/audio-page.tsx has 2366 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/audio/audio-page.tsx:1
• studio/frontend/src/features/export/export-page.tsx has 2197 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/export/export-page.tsx:1
• studio/frontend/src/features/hub/hub-page.tsx has 2197 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/hub/hub-page.tsx:1
• REDACTED has 2146 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — REDACTED:1
• studio/frontend/src/features/chat/chat-settings-sheet.tsx has 2107 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/chat-settings-sheet.tsx:1
• studio/frontend/src/features/chat/api/chat-api.ts has 2024 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/api/chat-api.ts:1
• studio/frontend/src/features/settings/tabs/voice-tab.tsx has 1900 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/settings/tabs/voice-tab.tsx:1
• studio/frontend/src/features/library/chats/chats-items.tsx has 1838 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/library/chats/chats-items.tsx:1
• studio/frontend/src/features/api-monitor/api-monitor-page.tsx has 1763 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/api-monitor/api-monitor-page.tsx:1
• studio/frontend/src/features/training/stores/training-config-store.ts has 1760 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/training/stores/training-config-store.ts:1
• studio/frontend/src/components/assistant-ui/code-fence-defer.tsx has 1741 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/components/assistant-ui/code-fence-defer.tsx:1
• studio/frontend/src/features/model-picker/components/model-selector/model-catalog.check.ts has 1688 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/model-picker/components/model-selector/model-catalog.check.ts:1
• studio/frontend/src/features/library/chats/chats-library.tsx has 1678 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/library/chats/chats-library.tsx:1
• studio/frontend/src/features/model-picker/components/model-selector/model-catalog.ts has 1656 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/model-picker/components/model-selector/model-catalog.ts:1
• studio/frontend/src/features/model-picker/model-config/per-model-config.ts has 1648 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/model-picker/model-config/per-model-config.ts:1
• REDACTED has 1629 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — REDACTED:1
• studio/frontend/src/components/assistant-ui/streaming-render-schedule.ts has 1608 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/components/assistant-ui/streaming-render-schedule.ts:1
• REDACTED has 1568 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — REDACTED:1
• studio/frontend/src/features/images/lib/image-metadata.ts has 1538 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/images/lib/image-metadata.ts:1
• studio/frontend/src/features/rag/components/thread-documents-bar.tsx has 1538 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/rag/components/thread-documents-bar.tsx:1
• studio/frontend/src/features/settings/tabs/agents-tab.tsx has 1516 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/settings/tabs/agents-tab.tsx:1
• studio/frontend/src/features/chat/provider-capabilities.ts has 1420 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/provider-capabilities.ts:1
• studio/frontend/src/features/chat/utils/chat-history-storage.ts has 1401 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/features/chat/utils/chat-history-storage.ts:1
• studio/frontend/src/lib/latex.ts has 1397 significant lines (blank lines excluded), over the 400-line bar, and contains logic — a file this long is usually several responsibilities in one module. — studio/frontend/src/lib/latex.ts:1

What to do

• Split each oversized file along the responsibilities already in it, into smaller focused modules in the same package.

React / JS · Readiness — Static test reachability (D-386): the share of production files reachable from any test via the import graph — measured without running anything.

Method: Static test reachability: the share of production files reachable from any test via the import graph — measured without running anything. Deterministic.

• No test imports this module directly or transitively. Import reachability cannot see a test that executes a file by path instead of importing it, nor one that drives it through a running browser by navigating to a URL — if neither does, no test reaches this one. (×40) — studio/frontend/smoke-heavy-thread-main.tsx:1, studio/backend/core/data_recipe/oxc-validator/validate.mjs:1, studio/frontend/smoke-code-block-flicker-main.tsx:1, …

What to do

• Add tests that import the unreached modules (directly or through their public entry).

React / JS · Readiness — How outdated the npm dependencies are (a maturity signal). JS/npm CVEs are scored separately in D30 (JS/npm Dependency Vulnerabilities).

Method: npm dependency staleness from manifest/registry metadata (a maturity signal; JS/npm CVEs are scored separately in D30, which answers dependency vulnerabilities for every ecosystem). Deterministic.

React / JS · Readiness — Whether the project wires up test, lint and typecheck — detected from each package.json script's COMMAND (eslint / tsc / vitest / jest / playwright), not just its name, and corroborated against CI-workflow invocations so a tool run only in CI still counts.

Method: package.json scanned for test/lint/typecheck script wiring. Deterministic presence check.

React / JS · Code Health — Files unreachable from every application/tooling/test entry point, and exports nothing imports (module-graph reachability, D-386).

Method: Dead code: files unreachable from every application/tooling/test entry point plus exports nothing imports, via module-graph reachability. Deterministic, exhaustive over the import graph.

• Unreachable from the 253 application, 12 tooling and 1460 test entry point(s) detected in this repo. Gate removals on `npm run build` — an undetected custom entry would make these reachable.
• no import path from any entry point (253 application, 12 tooling, 1460 test roots considered), and no other file in the scanned tree imports it — nothing in-repo names this module at all, which is the strongest form of this claim the import graph can make (×22) — studio/frontend/src/components/ui/menubar.tsx, studio/frontend/src/components/ui/navigation-menu.tsx, studio/frontend/src/components/ui/sparkles-text.tsx, …
• no import path from any entry point (253 application, 12 tooling, 1460 test roots considered), but 1 in-repo import(s) from 1 other file(s) do name it (studio/frontend/src/components/ui/toggle-group.tsx) — every one of those referrers is itself unreachable, so this file is dead only as a member of that cluster: if any referrer is in fact alive, this row falls with it — studio/frontend/src/components/ui/toggle.tsx
• no import path from any entry point (253 application, 12 tooling, 1460 test roots considered), but 1 in-repo import(s) from 1 other file(s) do name it (studio/frontend/src/components/layout/index.ts) — every one of those referrers is itself unreachable, so this file is dead only as a member of that cluster: if any referrer is in fact alive, this row falls with it (×2) — studio/frontend/src/components/layout/dashboard-grid.tsx, studio/frontend/src/components/layout/dashboard-layout.tsx
• Nothing imports this binding — it is safe to review for removal. (×14) — studio/frontend/src/components/assistant-ui/attachment.tsx:825, studio/frontend/src/components/assistant-ui/thread-width-steps.ts:4, studio/frontend/src/components/ui/animated-theme-toggler.tsx:95, …

What to do

• Delete the dead files and unused exports — every line is maintenance cost and rebuild-estimate inflation with zero runtime value.

React / JS · Readiness — npm dependency truthfulness (D-386): unused dependencies, imports not declared anywhere, and type-/test-only packages shipped as production deps.

Method: npm dependency truthfulness: unused dependencies, imports declared nowhere, and type-/test-only packages shipped as production deps — from the manifest + import graph. Deterministic.

• Imported but not declared in any reachable package.json — installs work only by hoisting accident. (×10) — studio/frontend/scripts/coal-span-census.mjs:37, studio/frontend/tests/attachment-preview-text.test.ts:7, studio/frontend/tests/math-block-marker.test.ts:57, …
• Declared in docker/jupyter/unsloth_labext/package.json but never imported anywhere in that package or its workspace members — no static import reaches it. Usually that is dead weight and attack surface, but two shapes are indistinguishable from source and are NOT dead: an optional or native peer that another dependency loads dynamically at runtime, and a package a build, docs or test step installs and invokes separately. Confirm which of the three this is before removing it. (×2)
• Declared in studio/frontend/package.json but never imported anywhere in that package or its workspace members — no static import reaches it. Usually that is dead weight and attack surface, but two shapes are indistinguishable from source and are NOT dead: an optional or native peer that another dependency loads dynamically at runtime, and a package a build, docs or test step installs and invokes separately. Confirm which of the three this is before removing it.
• Every import is type-only — move it to devDependencies. — studio/frontend/src/components/assistant-ui/code-fence-defer.tsx:6

What to do

• Remove unused dependencies, declare unlisted imports explicitly, and demote type-/test-only packages to devDependencies.

React / JS · Architecture — Import cycles in the module graph (D-386) — files that can only be understood and changed together.

Method: Import cycles in the module graph, detected exhaustively over JS/TS imports (the same cycle detection as the .NET coupling dimension). Deterministic.

• studio/frontend/src/features/chat/api/skills-api.ts → studio/frontend/src/features/chat/utils/refresh-context-usage.ts → studio/frontend/src/features/chat/api/chat-adapter.ts → studio/frontend/src/features/chat/api/skills-api.ts (one verified cycle inside a mutually-dependent group of 280 files) — studio/frontend/src/features/chat/api/skills-api.ts:8
• studio/frontend/src/features/chat/presets/preset-load-config.ts → studio/frontend/src/features/chat/stores/chat-runtime-store.ts → studio/frontend/src/features/chat/presets/preset-load-config.ts (one verified cycle inside a mutually-dependent group of 4 files) — studio/frontend/src/features/chat/presets/preset-load-config.ts:35
• studio/frontend/src/features/recipe-studio/blocks/registry.ts → studio/frontend/src/features/recipe-studio/blocks/render-dialog.tsx → studio/frontend/src/features/recipe-studio/dialogs/expression/expression-dialog.tsx → studio/frontend/src/features/recipe-studio/dialogs/shared/available-variables.tsx → studio/frontend/src/features/recipe-studio/stores/recipe-studio.ts → studio/frontend/src/features/recipe-studio/blocks/registry.ts (one verified cycle inside a mutually-dependent group of 9 files) — studio/frontend/src/features/recipe-studio/blocks/registry.ts:19
• studio/frontend/src/i18n/locale-store.ts → studio/frontend/src/i18n/messages.ts → studio/frontend/src/i18n/locale-store.ts — studio/frontend/src/i18n/locale-store.ts:5

What to do

• Break each cycle by extracting the shared piece into a module both sides can import.

Other · Code Health — Whether the code avoids sync-over-async (deadlock-prone blocking on tasks) and async void.

Method: Roslyn syntax scan: async methods scanned for .Wait()/.GetAwaiter().GetResult() and async-void outside event handlers. Deterministic, hard fact per invocation.

• The task this creates is not kept anywhere. The event loop holds only a weak reference to its tasks, so a task nobody references can be garbage-collected before it finishes (Python's documentation for `asyncio.create_task`: save a reference to the result); and nothing ever awaits it, so an exception it raises is never handled — it surfaces, if at all, as "Task exception was never retrieved" in a log. Keep the task: `await` it, gather it, start it in an `asyncio.TaskGroup` (`async with asyncio.TaskGroup() as tg: tg.create_task(…)`), or, for genuine fire-and-forget, add it to a set and `task.add_done_callback(tasks.discard)` with a callback that also reads its exception. (×2) — studio/backend/core/inference/openai_auto_download.py:1019, studio/backend/core/inference/chat_generation_runs.py:331

What to do

• Task from `asyncio.create_task` is not kept

Other · Code Health — Whether any branch is dead by construction — a switch arm whose label can never equal a case-normalised subject, or an `else if` whose predicate the arm above has already swallowed.

Method: Roslyn syntax + semantics: switch labels compared against the subject's own case normaliser, and if/else-if chains checked for a literal an earlier arm's containment test already swallows. Deterministic, provable per finding. Advisory.

Other · Code Health — Whether a child process that has BOTH standard streams redirected drains both — reading one to the end while the other is never read deadlocks once the child fills the unread pipe.

Method: Roslyn syntax + semantics: ProcessStartInfo launches with both streams redirected, checked for a drain of each stream across the enclosing type. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `spawn` imported from `child_process` whose stdout and stderr are both pipes (no options, no `stdio`, or `stdio` of `'pipe'`), bound to a local that never leaves its scope, where exactly one of the two streams is read, the other never, and the child’s `close`/`exit` (or the read stream’s end) is awaited; a shell redirect in the call’s arguments or a `kill` of the child suppresses it. The same rule reads Python’s `subprocess.Popen`, Dart’s `Process.start`, Ruby’s `Open3.popen3`, and Scala’s JDK process starts (`new ProcessBuilder(…).start()` with no `redirectErrorStream`/`redirectError`/`redirectOutput`/`inheritIO`, `Runtime.getRuntime.exec`) where one of `getInputStream`/`getErrorStream` is read and the other never while `waitFor()` is called or the read stream is read to its end. Deterministic, provable per finding. Advisory.

Other · Security — Whether a hand-rolled public/private IP check can be walked past — a method that unwraps IPv4-mapped IPv6 but returns the opposite verdict for the same host written as IPv4-compatible, 6to4 or NAT64.

Method: Roslyn syntax + semantics: methods that unwrap IPv4-mapped IPv6 and hand-roll IPv4 range carve-outs, checked for whether the IPv6 branch also accounts for the IPv4-compatible, 6to4 and NAT64 embeddings. Deterministic, provable per finding. Advisory.

Other · Security — Whether a length read out of the stream being parsed is bounded before it is allocated or read — an unchecked count taken from the input lets the input choose the allocation.

Method: Roslyn syntax + semantics: integer lengths read from a BinaryReader and spent on a bulk read or an array allocation, checked for any comparison or bounding call on the value anywhere in the method. Deterministic, provable per finding. Advisory.

• `size` is read from the bytes being parsed at line 420 (`struct.unpack("<Q")`), then passed as the size of `read` at line 421 — and nowhere in its function is `size` compared to anything, clamped, or bounded against the input that remains. Whoever wrote those bytes therefore chooses the allocation directly: the call reserves the whole length before a single byte of the payload is known to exist, so a 32-bit length asks for up to 4 GiB (a 64-bit one for more than any machine has), and the reservation either raises `MemoryError` out of a parser whose failure path was written for malformed input, not for hostile input, or succeeds and holds gigabytes for a few bytes on the wire. Bound `size` against the bytes actually remaining and a maximum the format allows before reading. — studio/backend/core/inference/video_ltx2.py:421
• `n` is read from the bytes being parsed at line 20242 (`struct.unpack("<Q")`), then passed as the size of `read` at line 20243 — and nowhere in its function is `n` compared to anything, clamped, or bounded against the input that remains. Whoever wrote those bytes therefore chooses the allocation directly: the call reserves the whole length before a single byte of the payload is known to exist, so a 32-bit length asks for up to 4 GiB (a 64-bit one for more than any machine has), and the reservation either raises `MemoryError` out of a parser whose failure path was written for malformed input, not for hostile input, or succeeds and holds gigabytes for a few bytes on the wire. Bound `n` against the bytes actually remaining and a maximum the format allows before reading. — studio/backend/core/inference/llama_cpp.py:20243
• `slen` is read from the bytes being parsed at line 20718 (`struct.unpack("<Q")`), then passed as the size of `read` at line 20719 — and nowhere in its function is `slen` compared to anything, clamped, or bounded against the input that remains. Whoever wrote those bytes therefore chooses the allocation directly: the call reserves the whole length before a single byte of the payload is known to exist, so a 32-bit length asks for up to 4 GiB (a 64-bit one for more than any machine has), and the reservation either raises `MemoryError` out of a parser whose failure path was written for malformed input, not for hostile input, or succeeds and holds gigabytes for a few bytes on the wire. Bound `slen` against the bytes actually remaining and a maximum the format allows before reading. — studio/backend/core/inference/llama_cpp.py:20719
• `n` is read from the bytes being parsed at line 3405 (`int.from_bytes(… 8 bytes …)`), then passed as the size of `read` at line 3406 — and nowhere in its function is `n` compared to anything, clamped, or bounded against the input that remains. Whoever wrote those bytes therefore chooses the allocation directly: the call reserves the whole length before a single byte of the payload is known to exist, so a 32-bit length asks for up to 4 GiB (a 64-bit one for more than any machine has), and the reservation either raises `MemoryError` out of a parser whose failure path was written for malformed input, not for hostile input, or succeeds and holds gigabytes for a few bytes on the wire. Bound `n` against the bytes actually remaining and a maximum the format allows before reading. — REDACTED:3406
• `size` is read from the bytes being parsed at line 4143 (`struct.unpack("<Q")`), then passed as the size of `read` at line 4144 — and nowhere in its function is `size` compared to anything, clamped, or bounded against the input that remains. Whoever wrote those bytes therefore chooses the allocation directly: the call reserves the whole length before a single byte of the payload is known to exist, so a 32-bit length asks for up to 4 GiB (a 64-bit one for more than any machine has), and the reservation either raises `MemoryError` out of a parser whose failure path was written for malformed input, not for hostile input, or succeeds and holds gigabytes for a few bytes on the wire. Bound `size` against the bytes actually remaining and a maximum the format allows before reading. — unsloth/models/loader_utils.py:4144

What to do

• Each finding names the read that takes the length from the input and the site that spends it, and states that the value is never compared in between. Confirm from the two, then bound the length against the remaining stream before using it.

Other · Code Health — Whether a loop that shortens a string until it fits a length budget has a floor — one with none grinds the value down to the empty string, or past it into a negative-length `Substring`.

Method: Roslyn syntax + semantics: while/do loops whose body's only effect on a string is to drop its last character, checked for whether anything — a direct comparison on the length, a body guard, a break — bounds that length below. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `while`/`do` loop whose body’s one assignment to a value is `x = x.slice(0, -1)` or `x = x.slice|substring|substr(0, x.length - 1)`, driven by a condition that reads `x.length` only as a term of a larger expression — never compared directly, never tested for truthiness, and with no other read of `x` — and whose body has no `break`, `return`, `throw` or `if` naming `x`. Deterministic, provable per finding. Advisory.

Other · Security — Whether a walk that recurses through a JSON/XML tree handed in by its caller bounds how deep it will go — an uncapped walk lets the document's nesting choose the stack depth, and the resulting StackOverflowException cannot be caught.

Method: Roslyn syntax + semantics: methods that take a JSON/XML document node and call themselves with a child of it, reachable from an externally-callable member of the same type that accepts a document, checked for any depth parameter, descent counter or threaded arithmetic anywhere in the walk. On a repository with no .NET source it reads TypeScript off the engine’s own token stream (test, vendored, generated and minified paths and `.d.ts` not) with the same rule: a function or method with a parameter typed as a document value — `unknown`, `any`, `object`, `Record<string, unknown|any>`, a JSON alias (`JsonValue`, `JSONObject`, …) or a DOM node global the file does not rebind — that calls itself (bare, or through `this` for a method, or hands itself to a call over the value as in `value.map(walk)`) with something it took out of that value and never through an ancestor accessor such as `closest()` or `parentElement`, that is exported or reached from an exported function (a public method of the same exported class) taking such a value, and that names no depth, level, nesting, recursion, remaining or budget anywhere and threads no `+`/`-` arithmetic through a self-call. Plain JavaScript is not read: with no annotation nothing tells a parsed document from a tree the code built itself. Deterministic, provable per finding. Advisory.

• `package_json_extra_refs` is public, takes a document from its caller (line 366) and hands it to `walk`, which calls itself at line 383 (`walk`) with a piece of `obj` (and at line 386) — and nowhere in `walk` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `walk` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — scripts/check_frontend_dep_removal.py:383
• `recipe_names_host_env_credential` is public, takes a document from its caller (line 617) and hands it to `_names_env_credential`, which calls itself at line 604 (`_names_env_credential`) with a piece of `value` (and at line 612) — and nowhere in `_names_env_credential` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_names_env_credential` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/data_recipe/service.py:604
• `to_jsonable` is public and takes `value: Any` from its caller, and it calls itself at line 122 (`to_jsonable`) with a piece of `value` (and at line 124) — and nowhere in `to_jsonable` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `to_jsonable` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/data_recipe/jsonable.py:122
• `to_preview_jsonable` is public and takes `value: Any` from its caller, and it calls itself at line 160 (`to_preview_jsonable`) with a piece of `value` (and at line 162) — and nowhere in `to_preview_jsonable` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `to_preview_jsonable` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/data_recipe/jsonable.py:160
• `to_preview_jsonable_row` is public and takes `row: Any` from its caller, and it calls itself at line 175 (`to_preview_jsonable_row`) with a piece of `row` — and nowhere in `to_preview_jsonable_row` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `to_preview_jsonable_row` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/data_recipe/jsonable.py:175
• `native_unflatten` is public, takes a document from its caller (line 348) and hands it to `_decode`, which calls itself at line 156 (`_decode`) with a piece of `value` (and at line 189) — and nowhere in `_decode` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_decode` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/prequant_native.py:156
• `normalize_function_schema` is public, takes a document from its caller (line 63) and hands it to `_normalize_schema_node`, which calls itself at line 33 (`_normalize_schema_node`) with a piece of `schema` (and at line 35, line 40, line 53) — and nowhere in `_normalize_schema_node` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_normalize_schema_node` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/openai_responses_shared.py:33
• `parse_mcp_config` is public, takes a document from its caller (line 148) and hands it to `_has_variable_reference`, which calls itself at line 44 (`_has_variable_reference`) with a piece of `value` (and at line 46) — and nowhere in `_has_variable_reference` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_has_variable_reference` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/mcp_config_import.py:44
• `relax_nested_object_key_order` is public, takes a document from its caller (line 175) and hands it to `_reorderable`, which calls itself at line 134 (`_reorderable`) with a piece of `schema` — and nowhere in `_reorderable` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_reorderable` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/llama_tool_schema.py:134
• `stream_chat_completion` is public, takes a document from its caller (line 1664) and hands it to `_sanitize_gemini_schema`, which calls itself at line 4900 (`_sanitize_gemini_schema`) with a piece of `node` (and at line 4924, line 4928, line 4944, line 4951, line 4963) — and nowhere in `_sanitize_gemini_schema` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_sanitize_gemini_schema` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/external_provider.py:4900
• `guided_runs_at_guidance_off_batch` is public, takes a document from its caller (line 470) and hands it to `_graph_handle_live`, which calls itself at line 143 (`_graph_handle_live`) with a piece of `handle` — and nowhere in `_graph_handle_live` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_graph_handle_live` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_zimage_cfg_split.py:143
• `probe_consumption` is public, takes a document from its caller (line 260) and hands it to `_collect_tensors`, which calls itself at line 191 (`_collect_tensors`) with a piece of `value` (and at line 194, line 196) — and nowhere in `_collect_tensors` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_collect_tensors` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_te_truncate.py:191
• `install` is public, takes a document from its caller (line 238) and hands it to `_plain`, which calls itself at line 78 (`_plain`) with a piece of `value` (and at line 80) — and nowhere in `_plain` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_plain` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_prompt_cache.py:78
• `put` is public, takes a document from its caller (line 148) and hands it to `_map_tensors`, which calls itself at line 109 (`_map_tensors`) with a piece of `obj` (and at line 111, line 113) — and nowhere in `_map_tensors` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_map_tensors` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_prompt_cache.py:109
• `packed_weight_fingerprint` is public, takes a document from its caller (line 1359) and hands it to `_hash_packed_payload`, which calls itself at line 1350 (`_hash_packed_payload`) with a piece of `tensor` — and nowhere in `_hash_packed_payload` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_hash_packed_payload` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — REDACTED:1350
• `onload` is public, takes a document from its caller (line 493) and hands it to `_copy_into`, which calls itself at line 259 (`_copy_into`) with a piece of `dst` — and nowhere in `_copy_into` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_copy_into` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_offload_prefetch.py:259
• `graph_key` is public and takes `obj: Any` from its caller, and it calls itself at line 704 (`graph_key`) with a piece of `obj` (and at line 706) — and nowhere in `graph_key` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `graph_key` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_cuda_graph.py:704
• `reset_all` is public, takes a document from its caller (line 2831) and hands it to `_holds_graphs`, which calls itself at line 2828 (`_holds_graphs`) with a piece of `handle` — and nowhere in `_holds_graphs` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_holds_graphs` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_cuda_graph.py:2828
• `load_comfy_quant_transformer` is public, takes a document from its caller (line 1099) and hands it to `take`, which calls itself at line 835 (`take`) with a piece of `value` (and at line 838) — and nowhere in `take` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `take` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_comfy_quant.py:835
• `graph_key` is public, takes a document from its caller (line 155) and hands it to `_walk`, which calls itself at line 128 (`_walk`) with a piece of `obj` (and at line 130, line 134) — and nowhere in `_walk` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_walk` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/diffusion_block_graph.py:128
• `compaction_receipt_field` is public and takes `value: Any` from its caller, and it calls itself at line 463 (`compaction_receipt_field`) with a piece of `value` — and nowhere in `compaction_receipt_field` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `compaction_receipt_field` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/context_window.py:463
• `normalize_navigation_url` is public and takes `raw: Any` from its caller, and it calls itself at line 507 (`normalize_navigation_url`) with a piece of `raw` — and nowhere in `normalize_navigation_url` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `normalize_navigation_url` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/core/inference/browser_agent.py:507
• `restore_inventory_handles` is public, takes a document from its caller (line 213) and hands it to `_restore`, which calls itself at line 224 (`_restore`) with a piece of `payload` (and at line 226, line 228) — and nowhere in `_restore` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_restore` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/hub/utils/host_paths.py:224
• `redact_host_paths` is public, takes a document from its caller (line 284) and hands it to `_redact`, which calls itself at line 337 (`_redact`) with a piece of `payload` (and at line 409, line 413, line 420) — and nowhere in `_redact` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_redact` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/hub/utils/host_paths.py:337
• `response_leaks_host_path` is public, takes a document from its caller (line 311) and hands it to `_find_leak`, which calls itself at line 444 (`_find_leak`) with a piece of `payload` (and at line 470, line 476) — and nowhere in `_find_leak` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_find_leak` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/hub/utils/host_paths.py:444
• `redact_inventory_error_detail` is public and takes `detail: Any` from its caller, and it calls itself at line 581 (`redact_inventory_error_detail`) with a piece of `detail` (and at line 586) — and nowhere in `redact_inventory_error_detail` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `redact_inventory_error_detail` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/hub/utils/host_paths.py:581
• `upsert_chat_settings_merge` is public, takes a document from its caller (line 5755) and hands it to `_deep_merge_settings`, which calls itself at line 5724 (`_deep_merge_settings`) with a piece of `current` — and nowhere in `_deep_merge_settings` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_deep_merge_settings` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/storage/studio_db.py:5724
• `upsert_chat_settings_merge_if_current` is public, takes a document from its caller (line 5822) and hands it to `_chat_settings_match_expected`, which calls itself at line 5806 (`_chat_settings_match_expected`) with a piece of `current` — and nowhere in `_chat_settings_match_expected` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_chat_settings_match_expected` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/storage/studio_db.py:5806
• `from_route` is public, takes a document from its caller (line 21) and hands it to `_unwrap`, which calls itself at line 41 (`_unwrap`) with a piece of `annotation` — and nowhere in `_unwrap` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `_unwrap` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/studio_mcp/outputs.py:41
• `drop_non_finite` is public and takes `value: Any` from its caller, and it calls itself at line 179 (`drop_non_finite`) with a piece of `value` (and at line 181) — and nowhere in `drop_non_finite` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `drop_non_finite` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/backend/utils/training_runs.py:179
• `mergedToolCallArgumentsText` is exported, takes a document from its caller (line 381) and hands it to `mergeJsonNode`, which calls itself at line 354 (`mergeJsonNode`) with a piece of `current` (and at line 366) — and nowhere in `mergeJsonNode` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. JavaScript sets no limit of its own on that nesting: `JSON.parse` accepts any depth, so a 100,000-byte body — inside Express's default 100kb limit — parses to 50,000 levels, and a walk like this one throws `RangeError: Maximum call stack size exceeded` long before it reaches the bottom. Wherever that throw lands, the work fails: a request errors, and a walk running outside any `try` — an event handler, a timer, a promise nobody awaits — takes the Node process down with every other request in flight. Thread a depth argument through `mergeJsonNode` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/frontend/src/features/chat/tool-call-arguments.ts:354
• `extractDeltaText` is exported, takes a document from its caller (line 14) and hands it to `extractReasoningText`, which calls itself at line 22 (`extractReasoningText`) with a piece of `payload` (and at line 29) — and nowhere in `extractReasoningText` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. JavaScript sets no limit of its own on that nesting: `JSON.parse` accepts any depth, so a 100,000-byte body — inside Express's default 100kb limit — parses to 50,000 levels, and a walk like this one throws `RangeError: Maximum call stack size exceeded` long before it reaches the bottom. Wherever that throw lands, the work fails: a request errors, and a walk running outside any `try` — an event handler, a timer, a promise nobody awaits — takes the Node process down with every other request in flight. Thread a depth argument through `extractReasoningText` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/frontend/src/features/chat/utils/parse-assistant-content.ts:22
• `chatContentPartAttachmentSignature` is exported, takes a document from its caller (line 46) and hands it to `stableJson`, which calls itself at line 31 (`stableJson`) with a piece of `value` (and at line 39) — and nowhere in `stableJson` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. JavaScript sets no limit of its own on that nesting: `JSON.parse` accepts any depth, so a 100,000-byte body — inside Express's default 100kb limit — parses to 50,000 levels, and a walk like this one throws `RangeError: Maximum call stack size exceeded` long before it reaches the bottom. Wherever that throw lands, the work fails: a request errors, and a walk running outside any `try` — an event handler, a timer, a promise nobody awaits — takes the Node process down with every other request in flight. Thread a depth argument through `stableJson` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — studio/frontend/src/features/chat/utils/chat-attachment-events.ts:31

What to do

• Each finding names the externally-callable member that accepts the document and the self-call that descends, and states that no depth parameter or counter exists anywhere in the walk. Confirm from the two, then thread a depth argument through the walk and reject documents past a fixed maximum, or rewrite it with an explicit stack.

Other · Code Health — Whether a type's disposal matches what it OWNS — releasing what it created, leaving alone what it was handed, and not declaring a finalizer for state that has nothing unmanaged to finalize.

Method: Roslyn syntax + semantics: every assignment to a disposable field is read to decide whether the type CREATED the value or was handed it, and the type's disposal is checked against that answer — an injected interface it disposes, a value it constructed and never releases, a finalizer on a type holding nothing unmanaged, and a disposable local whose every reference is a plain member read. A value handed to a container that disposes its contents (a parent control's `Controls` collection, a component `IContainer`) is released by that container and is not reported; generated code is out of population. On a repository with no .NET source the same ownership questions are read in JavaScript/TypeScript off the engine’s own token stream (test, vendored, generated and minified paths not): a class declaring `dispose()`, `[Symbol.dispose]()` or `[Symbol.asyncDispose]()` that disposes a field it was handed through a constructor parameter typed as a repository interface or resolved by a dependency-injection container; that assigns a field only ever from `new X(…)` of a disposable class and neither releases it anywhere in the class nor names it in its disposal member or a method that member calls, nor hands it to anything else; and a `const`/`let` local built from literals only whose every reference opens a statement operating on a non-release member of it. A class is disposable when every repository declaration of its name declares or inherits a disposal member, or when it is a documented library disposable (`vscode` EventEmitter, CancellationTokenSource and Disposable; three.js geometries, materials, textures, render targets, renderers, controls and composers). The finalizer arm has no JavaScript counterpart: a class cannot declare one. Dart is read for the owned-and-never-released arm: a Flutter `State` (or a class declaring `dispose()`/`close()`) whose library-private field is written only with direct creations of a framework controller, `FocusNode`, `StreamController` or periodic `Timer` (or a `.listen(…)` into a field declared `StreamSubscription`), which nothing in the class or its library releases (`dispose`/`close`/`cancel`) and which is never handed on except as a named argument. Ruby is read for the owned-and-never-released arm and the abandoned-local arm: a class declaring an instance `close`, `dispose`, `shutdown`, `stop`, `disconnect` or `terminate` whose instance variable is written only with a blockless core descriptor creation (`File.open`/`new`, `Tempfile`, `Dir.open`, `IO.popen`, the `socket` and `openssl` socket classes) that no method of the class closes and that is never handed on (passed, returned, stored, exposed by `attr_reader`, delegated) — a class whose superclass or mixin lies outside the repository is not judged — and a blockless creation read through and dropped (`File.open(path).read`) or held in a local that is never closed, returned, passed or stored (a `&:read` block-pass, an `autoclose: false` borrow and a `fileno` hand-over are not abandonment). PHP is read for the same two arms: a class (not a trait, and not one using a trait) declaring an instance `close`, `dispose`, `shutdown`, `stop`, `disconnect` or `terminate` whose `private` property (or a `final` class's `protected` one) is written only with a direct core `fopen`, `tmpfile`, `popen`, `proc_open`, `fsockopen`, `stream_socket_client`/`server` or `opendir` that nothing in the class closes (`fclose`, `pclose`, `closedir`, `unset`, a reset to `null`) and that is never handed on, and a handle held in a function's local, or read through as the argument of a reader, that is never closed, returned, passed, captured or stored; a `php://` stream, `curl_init`, `SplFileObject`, `PDO` and a function the repository declares under the core name are not read. Deterministic, provable per finding. Advisory.

Other · Code Health — Whether a method that temporarily changes state belonging to the whole process — the working directory, an environment variable — puts it back on EVERY path: a restore reached only when nothing throws leaks the change to the rest of the process.

Method: Roslyn syntax + semantics: method bodies that write the process working directory or an environment variable and write it back in the same body, checked for whether that restore sits in a `finally`/`catch` or only on the straight-line path. On a repository with no .NET source the same rule reads production JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not): `process.chdir`, `process.env.NAME =`/`["NAME"] =` and `delete process.env.NAME`, and Deno’s `Deno.chdir`/`Deno.env.set`/`Deno.env.delete`, paired per function body (a nested function or arrow is its own body, and module top-level code is none), where the last write puts back a local the body captured from the same global or deletes a variable the first write set, with at least one statement between them; a write in a `catch`/`finally` of that body silences it. Deterministic, provable per finding. Advisory.

• `_switch_to_bundled_tcc` changes state that belongs to the whole PROCESS and changes it back at the end of the same body — the `CC` environment variable (set at line 204, cleared at line 210) — which says the change was meant to last no longer than this call. But the restore is an ordinary statement on the straight-line path, not a `finally`, so it is skipped by any exception raised in between: the process then keeps the overwritten value for the rest of its life, and every later relative path, `subprocess` launch, `os.environ` read or `print` in ANY module sees it. That failure is silent and remote from its cause — the function that breaks is not the function that moved the value. The window is not private while it is open either: the working directory, the environment, the umask and `sys.stdout` are per-process, not per-thread, so a concurrent thread sees the moved value too. The durable fix is not to widen the restore but to stop mutating the global: pass the value to the operation that needs it — the `cwd=` and `env=` arguments of `subprocess.run`/`Popen` scope both to the CHILD, absolute paths built with `os.path.join` or `pathlib` need no working directory, and a stream handed to the code that writes needs no `sys.stdout`. Where the global genuinely must move, `contextlib.chdir` (3.11+), `contextlib.redirect_stdout` or a `try`/`finally` makes the restore run on every path. — studio/backend/core/_msvc_env.py:195

What to do

• Each finding names the statement that changes the process-global value and the statement that puts it back. Confirm from the two that nothing between them is protected — then prefer removing the mutation over widening the restore: a working directory or an environment variable passed to the operation that needs it (the `cwd` and `env` options of `child_process`, or of `Deno.Command`) is scoped to that operation, which is what the save-and-restore was trying to approximate. Where the global genuinely must move, a `try`/`finally` is the least that makes the restore unconditional. In Python the same holds for `os.chdir`, `os.environ` and `sys.stdout`: pass `cwd=`/`env=` to `subprocess`, or use `contextlib.chdir`/`redirect_stdout`, rather than moving the global by hand.