Lines 3421-3461markdown
3421 D005 C# vulnerabilities defect: parser written against synthetic
3422 output, never validated against real tool output. Fixed by
3423 matching the canonical `\): error \w+:` regex. Caught by adding
3424 the C# row to the new lint matrix; the row failed because the
3425 parser returned 0 despite exit code != 0 and visible violations
3426 in the output. (`src/languages/csharp.ts`, **D016**)
3430- **Cross-ecosystem matrix layer** (`test/integration/cross-ecosystem.test.ts`).
3431 New `BENCHMARK_LANGUAGES` table at the top of the file is the
3432 single source of truth for which languages participate and where
3433 each fixture's deliberate findings live. Each `describe('matrix —
3434 <report>')` block iterates the table to produce one uniform
3435 assertion per language — adding a new feature is one new
3436 optional field per row + one new `matrix —` describe; adding a
3437 6th language is one row append + one fixture dir + one CI install.
3438 No search-and-replace across describe blocks.
3440- **`matrix — secrets` (Phase 10i.0.1)** — 5 hardcoded fake AWS
3441 access keys (`AKIA1234567890ABCDEF` — patterned digits/letters
3442 that pass gitleaks' `aws-access-token` regex but fail real AWS
3443 validation and GitHub push protection). One per benchmark
3444 ecosystem. Asserts `dxkit vulnerabilities` surfaces a
3445 `SecretFinding` (category=secret, tool=gitleaks,
3446 rule=aws-access-token) for each.
3448- **`matrix — lint` (Phase 10i.0.2)** — 5 deliberate idiomatic
3449 linter violations (Python ruff F401 unused-import, Go gosimple
3450 S1002 bool-comparison, Rust clippy unused_variables, C#
3451 dotnet-format whitespace × 2). Asserts `dxkit quality` reports
3452 the expected linter and ≥1 lint finding. CI workflow now
3453 installs `ruff` (pipx), `golangci-lint` (curl install script),
3454 and `clippy` (rustup component) alongside the existing depVulns
3455 toolchains; `dotnet format` ships in the .NET 8 SDK.
3457- **`matrix — duplications` (Phase 10i.0.3)** — two near-identical
3458 helpers per fixture, sized comfortably above jscpd's
3459 `--min-lines 5 --min-tokens 50` defaults (initial pass had
3460 ~30-token bodies that fell below the threshold; widened on the
3461 way in). Asserts `metrics.duplication.cloneCount > 0`.