[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77K70RH9ZG4W09085VT3F1Y":3,"skill-same:01M77K70RH9ZG4W09085VT3F1Y":87},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":33,"commitSha":34,"skillDirectory":35,"forksCount":36,"repositoryHtmlUrl":37,"repositoryHomepageUrl":6,"ownerHtmlUrl":38,"ownerBlogUrl":39,"ownerDescription":40,"updatedAt":41,"install":42,"resources":54},"01M77K70RH9ZG4W09085VT3F1Y","audit-coverage",null,"Read coverage artifacts a build already wrote (lcov, Cobertura, coverage.py JSON, a Go cover profile) and report line coverage per file and function, plus CRAP per function, for a change, a path, or the tree. Runs no tests, never gates. Use when: 'coverage of this change', 'coverage per function', 'CRAP score', 'which functions are complex and untested', 'read the lcov report', 'coverage.xml', 'how covered is this file'. Complexity alone: \u002Fcode-metrics:audit complexity.","melodic-software\u002Fclaude-code-plugins","melodic-software","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F58273638?v=4","01M77JVQTXNGKCE4MDFY2BC936","Melodic Software","01M77JVQVN9QDWKAH6W03ANP76",30,"MIT","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\ndescription: \"Read coverage artifacts a build already wrote (lcov, Cobertura, coverage.py JSON, a Go cover profile) and report line coverage per file and function, plus CRAP per function, for a change, a path, or the tree. Runs no tests, never gates. Use when: 'coverage of this change', 'coverage per function', 'CRAP score', 'which functions are complex and untested', 'read the lcov report', 'coverage.xml', 'how covered is this file'. Complexity alone: \u002Fcode-metrics:audit complexity.\"\nargument-hint: \"[--json] [--all] [--base \u003Cref>] [--artifacts \u003Cpath>]... [\u003Cpath>...]\"\nuser-invocable: true\ndisable-model-invocation: false\nallowed-tools: [\"Bash(${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh:*)\", \"Bash(git branch --show-current:*)\"]\nshell: bash\nmetadata:\n  workflow-stage: anytime\n  summary: Coverage and CRAP read from build artifacts, no verdict\n---\n\n## Pre-computed context\n\nCurrent branch: !`git branch --show-current 2>\u002Fdev\u002Fnull || echo \"unknown\"`\n\n## Purpose\n\nCoverage is a dynamic measure. This skill reads it statically: it parses an artifact a test run\nalready wrote and joins it to the functions in scope. Running the tests is the project's own job\nand stays there, so nothing here executes a test command, starts a runner, or installs a tool.\n\nTwo numbers come out. Line coverage per file and per function is read from the artifact. CRAP per\nfunction is derived: `comp^2 * (1 - cov\u002F100)^3 + comp`, where `comp` is the function's cyclomatic\ncomplexity from the `audit` skill's `audit-complexity.sh` script, run over the same scope.\n\n| Format | Written by | What the join gets from it |\n|---|---|---|\n| lcov `.info` | most JavaScript and C\u002FC++ toolchains, `coverage lcov` | line hits; function records in either lcov record shape, which is where a function end line can come from at all |\n| Cobertura XML | gcovr, coverlet, kcov (which is how a Bash lane gets an artifact) | line hits, `\u003Csources>` prefixes, and `\u003Cmethod>` regions with their hit flag |\n| coverage.py JSON | `coverage json` | executed and missing lines, and per-function regions |\n| Go cover profile | `go test -coverprofile` | statement counts over line ranges, which is not a line table: the profile never says which lines carry the statements, so a Go file reports the statement ratio `go tool cover -func` prints, labels it `cov_source: statement-ratio`, and reports `lines_executable` and `lines_hit` as null. The format names no functions either, so a Go function reports `coverage_pct` and `crap` as null, and its line counts as null unless a line-measuring artifact also covered the file |\n\nThe record shape and version behind each format are its row under \"Coverage artifacts\" in\n`${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fcollectors.md`, which carries each claim with its basis, date and\nrecheck trigger.\n\nThe SQLite data file coverage.py writes while measuring is never read; why is the coverage.py\nSQLite row of that file. Run `coverage json` or `coverage lcov` and point this skill at the result.\n\n## Run it\n\n```bash\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\"                                  # the change, with the artifact auto-discovered\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --artifacts build\u002Flcov.info src\u002F # an explicit artifact and an explicit scope\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --all --artifacts coverage.xml --artifacts cover.out\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --json --all                     # the code-metrics\u002Fv2 document instead of markdown\n```\n\nWith no `--artifacts` and no `coverage.artifacts` in the config, the well-known names are looked\nfor under the repository root, at most two directory levels deep and never inside `node_modules`,\n`.git` or `vendor`: `coverage\u002Flcov.info`, `lcov.info`, `coverage.xml`, `cobertura.xml`,\n`coverage.json`, `coverage.out`, `cover.out`. Each artifact's format is detected from its content,\nnot its extension.\n\nPresent the markdown report as printed: the scope, the \"Coverage of this run\" table, the\nreferences with their provenance, then one row per file and per function. Keep the `--json`\ndocument when the numbers feed a comparison: pass it to `\u002Fverification:measure metrics` when the\n`verification` plugin is installed, treating `status: empty` on either side as INCONCLUSIVE;\notherwise keep the JSON beside your notes and compare by hand.\n\n## Getting a first artifact\n\nA run that finds no artifact reports every lane `unavailable`, lists the paths it searched, and\nends its markdown with a line pointing here, because the remedy is the project's to apply: this\nskill runs no test and installs nothing. One command per lane produces an artifact the next run\ncan read. Each row restates a producer's documented default, verified against the page in its\nBasis column on 2026-09-12; the producer's page wins over the row. Recheck trigger: a release note\nor changelog entry of that producer naming the flag or the output path in its row, or a read-time\nfetch of the Basis page that no longer states what the row does; either re-derives the row from the\npage and refreshes the date.\n\n| Lane | Producer | Command shape | Writes | Basis |\n|---|---|---|---|---|\n| TypeScript\u002FJavaScript | vitest | `npx vitest run --coverage --coverage.reporter=lcov` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered | [Vitest coverage guide](https:\u002F\u002Fvitest.dev\u002Fguide\u002Fcoverage.html), `coverage.reporter` |\n| TypeScript\u002FJavaScript | jest | `npx jest --coverage` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered; the default `coverageReporters` list carries `lcov` | [Jest CLI](https:\u002F\u002Fjestjs.io\u002Fdocs\u002Fcli), `--coverage`; [Jest configuration](https:\u002F\u002Fjestjs.io\u002Fdocs\u002Fconfiguration), `coverageReporters` default `[\"clover\", \"json\", \"lcov\", \"text\"]` |\n| TypeScript\u002FJavaScript | c8 or nyc, over any test runner | `npx c8 --reporter=lcov \u003Ctest command>` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered | [c8 README](https:\u002F\u002Fgithub.com\u002Fbcoe\u002Fc8), `--reporter`, which takes any [Istanbul reporter](https:\u002F\u002Fistanbul.js.org\u002Fdocs\u002Fadvanced\u002Falternative-reporters\u002F), `lcov` among them |\n| Python | coverage.py | `python -m coverage run -m pytest && python -m coverage json` | coverage.py JSON at `coverage.json`, auto-discovered; `coverage xml` writes `coverage.xml`, also auto-discovered; `coverage lcov` writes `coverage.lcov`, which needs `--artifacts` | coverage.py command pages [`run`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_run.html), [`json`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_json.html), [`xml`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_xml.html), [`lcov`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_lcov.html) |\n| Python | pytest-cov | `pytest --cov=\u003Cpackage> --cov-report=json` | coverage.py JSON at `coverage.json`, coverage.py's default name for the `json` report, auto-discovered; `--cov-report=json:\u003Cpath>` moves it | [pytest-cov reporting](https:\u002F\u002Fpytest-cov.readthedocs.io\u002Fen\u002Flatest\u002Freporting.html), `--cov-report`; the default file name is coverage.py's, per its [`json`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_json.html) page |\n| Bash | kcov | `kcov \u003Coutdir> bash \u003Ctest script>` | Cobertura-compatible XML under `\u003Coutdir>`; pass the `cobertura.xml` it writes with `--artifacts` | [kcov README](https:\u002F\u002Fgithub.com\u002FSimonKagstrom\u002Fkcov), \"Kcov will also write cobertura-compatible XML output\" |\n| Go | `go test` | `go test .\u002F... -coverprofile=coverage.out` | Go cover profile at the path `-coverprofile` names; `coverage.out` and `cover.out` are auto-discovered; file rows carry the statement ratio, function rows need a line artifact too | [`go` command, testing flags](https:\u002F\u002Fpkg.go.dev\u002Fcmd\u002Fgo#hdr-Testing_flags), `-coverprofile` |\n| C# | coverlet, as the `dotnet test` collector | `dotnet test --collect:\"XPlat Code Coverage\"` | Cobertura XML at `TestResults\u002F\u003Crun id>\u002Fcoverage.cobertura.xml`; pass it with `--artifacts`. The C# complexity lane is deferred, so the lane reports file rows and no CRAP | [coverlet VSTest integration](https:\u002F\u002Fgithub.com\u002Fcoverlet-coverage\u002Fcoverlet\u002Fblob\u002Fmaster\u002FDocumentation\u002FVSTestIntegration.md), `--collect:\"XPlat Code Coverage\"` |\n\nAuto-discovered means the output lands on one of the well-known names the run looks for with no\n`--artifacts`; anything else is named explicitly or listed under `coverage.artifacts` in the\nconfiguration.\n\n## Reading the numbers\n\n- A file row is the artifact's own line table: `coverage_pct`, `lines_executable`, `lines_hit`.\n  A file a Go cover profile covers is the exception: the profile weighs statements, so the\n  percentage is its statement ratio, `cov_source` reads `statement-ratio`, and both line counts\n  are null.\n- A function row exists for every function whose complexity collector reported a real end line.\n  It carries `coverage_pct`, `cyclomatic`, `crap`, and, in the JSON, `cov_source` and `hit`.\n- The markdown table is ordered so the rows worth reading survive its 200-row cap: rows over a\n  reference first, then function rows by CRAP descending with a null CRAP last, then file rows by\n  coverage ascending with a null percentage last, and file and line only as the tie-break. The\n  first function row is the highest-CRAP function in scope, however its file name sorts; the rows\n  the cap drops are counted on the table's last row and are all in the `--json` document, which is not\n  capped.\n- `cov_source` says where the coverage came from: `artifact-region` when the artifact carried the\n  function's own region (coverage.py `functions`, a Cobertura `\u003Cmethod>`, an lcov 2.2 `FNL` end\n  line), `line-range` when the range came from the complexity collector, in which case nested\n  function ranges are subtracted from the parent first, `statement-ratio` on a file row whose\n  percentage came from a Go profile's statements rather than from a line table, and `ambiguous`\n  on a function the join refused. The markdown table renders the values; the per-row `cov_source`\n  and `hit` fields are in the `--json` document.\n- A function the artifact names by a short name binds to the function whose line range holds the\n  lines the artifact recorded, not to whichever function the name happens to end at. Two `run`\n  methods in one file and a record named `run` are one function's coverage, and the range says\n  whose. Where nothing separates the candidates, because the artifact placed none of them at a\n  line or because two of them fall inside the same range, the join is refused rather than guessed:\n  the row reads `cov_source: ambiguous`; coverage, line counts, hit, and CRAP are null because\n  those numbers would be a guess, while cyclomatic from the complexity row is kept because that\n  collector already measured it; it carries a `coverage-ambiguous` label, and the lane's run row\n  turns `partial` and names the functions it left unjoined. A missing coverage number an operator\n  can see beats a wrong number they cannot.\n- `hit` is the artifact's function-hit flag. When it says the function was never entered,\n  `coverage_pct` is 0 rather than the 1\u002FN a declaration line executed at import would produce.\n- A function with no executable lines in the artifact reports `coverage_pct: null` and\n  `crap: null`, never 0: a fabricated 0 percent would produce a fabricated maximal CRAP. Its\n  `lines_executable` and `lines_hit` are 0 when an artifact that measures lines covers the file\n  and carries none in that range, and null when no artifact covering the file measures lines at\n  all, which is every function of a file only a Go profile covers.\n- CRAP is Savoia and Evans' formula (Agitar Labs, 2007). It is a description of two numbers you\n  already have, not a validated predictor of change risk, and no standard sets a threshold for it.\n  `\u002Fcode-metrics:principles` carries the provenance, the rename history, and the caveats; read it\n  before you quote a CRAP number at anyone.\n- A lane whose resolved complexity collector reports no function end lines gets a `\u003Clane>\u002Fcrap`\n  row with `status: not-applicable` and that reason. Bash is that lane in this version: no\n  maintained Bash collector reports an end line, so Bash CRAP is a stated gap, not a null.\n- A lane whose cyclomatic collector did not resolve, or ran and produced nothing parseable, gets a\n  `\u003Clane>\u002Fcrap` row with `status: unavailable` naming that collector and its reason, whatever the\n  lane's coverage row says. CRAP needs both numbers, and the missing one is the cause worth\n  reading; when the coverage row is not `ok` either, its reason follows the collector's on the\n  same row, so a partial count or a no-artifact search never hides a collector failure and the\n  collector never hides them. `Functions: 0` in the summary reads with that row.\n- A lane matched by fewer than all of its scope files carries `status: partial` and the reason\n  `partial, N of M scope files present in the artifacts`, so a total miss never reads as \"no\n  executable lines\" and the document cannot settle as `complete` while a row says `N of M`. A lane\n  no artifact covers is `unavailable` with that count; when no artifact was found at all, the\n  reason lists every path searched. A lane that left a function unjoined is `partial` for that\n  reason too.\n- A row carrying that `N of M` count also says which files: its reason names the first five scope\n  files no artifact mentions (`; missing: a, b, c, d, e, +N more in the JSON`) and the JSON row's\n  `missing` key lists all of them, root-relative and sorted. The count alone cannot tell a test\n  file an artifact never records from a source file the suites never reach; the paths can.\n- Whether the covered code is actually checked by its tests is a different question, and coverage\n  alone cannot answer it: a line can execute under a test that asserts nothing.\n  `\u002Fmutation-testing:audit` owns that question when the `mutation-testing` plugin is installed;\n  without it, say that the coverage number is an execution count and stop there.\n- Exit 0 whenever a report was produced, including an empty one; exit 2 for a usage error, which\n  includes a named artifact or scope path that does not exist; exit 3 when a complexity collector\n  ran and produced nothing parseable. On exit 3 the collector's stderr is on the `\u003Clane>\u002Fcrap`\n  row, and the summary carries an `Exit 3:` line naming the lane and collector, so the markdown\n  alone says why the run has no function rows.\n\n## Configuration\n\nEverything tunable resolves through the code-metrics layers (user-global, team\n`docs\u002Fconventions\u002Fcode-metrics.yaml`, personal override; per-key override; keys in `${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fconfig.md`):\n\n| Key | Default | Effect |\n|---|---|---|\n| `coverage.artifacts` | `[]` | Explicit artifact paths; empty means auto-discover. A listed path that does not exist is a usage error, so a stale entry is loud |\n| `coverage.path_prefix_strip` | `[]` | Prefixes removed from artifact paths before the join, for compiled-output layouts (`dist\u002F`, `build\u002F`) |\n| `coverage.reference` | `null` | No bar. A number counts rows below it as `over_reference` and nothing else |\n| `coverage.crap.reference` | `null` | No bar; no standard sets one for CRAP |\n\n`\u002Fcode-metrics:setup` writes the team file and probes the collectors.\n\n## What this skill does not do\n\n- It never runs tests, starts a test runner, or installs a coverage tool. It reads what a build\n  already wrote; if there is no artifact, the report says so and stops.\n- It never reads the SQLite data file coverage.py writes while measuring.\n- It never judges. A reference here is a value to count against, and no finding, severity, or\n  exit-code gate follows from one.\n- It does not measure branch, condition, or MC\u002FDC coverage. Those records are parsed past, not\n  reported, because the join is a line-level one.\n\n## Next\n\n- The numbers feed a before-and-after comparison: `\u002Fverification:measure metrics`.\n- Coverage looks healthy but the tests may assert nothing: `\u002Fmutation-testing:audit`.\n- A number is about to be quoted at someone: `\u002Fcode-metrics:principles`.\n\n## Gotchas\n\n- The complexity collector decides which functions can carry CRAP at all. With no collector\n  resolving for a lane, that lane has no function rows and its `\u003Clane>\u002Fcrap` row is `unavailable`\n  naming the collector's own reason. File-level coverage is unaffected: the scope the join keys on\n  is the one the audits measure, not the files a complexity collector emitted rows for, so a file\n  with no functions still reports its lines.\n- Path shapes differ between artifact and source tree. The join normalizes both sides, strips the\n  repository root and the configured prefixes, then falls back to a component-wise suffix match\n  and finally to a unique basename (which is how a Go profile's module path resolves). Both\n  fallbacks require a single winner: an artifact path that fits two scoped files equally well\n  (two services vendoring the same `pkg\u002Fa.py`) stays unmatched and shows up in the partial count,\n  because attributing it to either would credit one service with the other's coverage. The\n  basename fallback needs a single winner on the artifact side too, counted across every artifact\n  read rather than within one: when two services each ship a `handler.go` and only one is in\n  scope, the out-of-scope path stays unmatched instead of merging its coverage into the scoped\n  file, even though the scope names exactly one candidate.\n- An artifact older than the source it describes joins by line number and will be wrong without\n  saying so. Re-run the tests before reading the numbers when the diff has moved lines.\n- Two artifacts covering one file are merged by keeping the larger hit count per line, so a line\n  is never counted twice. Two artifacts covering one function are folded the same way: the hit\n  flag is the larger of the two, so a suite that never entered the function cannot report it at 0\n  percent beside a file row the other suite already showed as covered. Two Go profiles covering\n  one file merge block by block on the profile's own `start.col,end.col` identity and are totaled\n  after that, so two test shards that entered different blocks report what they covered between\n  them rather than what the luckier shard covered alone.\n- A Go profile and a line artifact naming the same `.go` file do not average or overwrite each\n  other: the profile's statement ratio is that file's exact native measure and wins the\n  percentage, the row says `statement-ratio`, and the line counts stay null rather than reading as\n  the counts behind a number they did not produce. The lane's `collector` still names both formats\n  read. Function rows are line-based, so a Go function does take its region from a line artifact\n  when one covers it.\n- A Cobertura report with several `\u003Csource>` roots is read against all of them: a relative class\n  filename takes the first root under which that path exists in the scanned tree, and falls back\n  to the first root when no candidate exists. Two roots that both hold the file is an ambiguity\n  the report cannot settle, so the first one wins and the class lands under it. The probe settles\n  anything only when the roots are relative, or absolute and present on the machine running the\n  audit. A report built elsewhere declares absolute roots (the usual coverlet, kcov and gcovr shape\n  for a CI run) naming directories this machine does not have, so every candidate misses and every\n  class lands under the first root. Rewriting a root from another machine onto the local tree\n  needs a mapping the report does not carry.\n- Two functions in one file whose qualified names share a tail (`A.run` and `B.run`) can bind to\n  each other's artifact region when the artifact carries only the short name; the file-level\n  numbers are unaffected.\n","## Pre-computed context\n\nCurrent branch: !`git branch --show-current 2>\u002Fdev\u002Fnull || echo \"unknown\"`\n\n## Purpose\n\nCoverage is a dynamic measure. This skill reads it statically: it parses an artifact a test run\nalready wrote and joins it to the functions in scope. Running the tests is the project's own job\nand stays there, so nothing here executes a test command, starts a runner, or installs a tool.\n\nTwo numbers come out. Line coverage per file and per function is read from the artifact. CRAP per\nfunction is derived: `comp^2 * (1 - cov\u002F100)^3 + comp`, where `comp` is the function's cyclomatic\ncomplexity from the `audit` skill's `audit-complexity.sh` script, run over the same scope.\n\n| Format | Written by | What the join gets from it |\n|---|---|---|\n| lcov `.info` | most JavaScript and C\u002FC++ toolchains, `coverage lcov` | line hits; function records in either lcov record shape, which is where a function end line can come from at all |\n| Cobertura XML | gcovr, coverlet, kcov (which is how a Bash lane gets an artifact) | line hits, `\u003Csources>` prefixes, and `\u003Cmethod>` regions with their hit flag |\n| coverage.py JSON | `coverage json` | executed and missing lines, and per-function regions |\n| Go cover profile | `go test -coverprofile` | statement counts over line ranges, which is not a line table: the profile never says which lines carry the statements, so a Go file reports the statement ratio `go tool cover -func` prints, labels it `cov_source: statement-ratio`, and reports `lines_executable` and `lines_hit` as null. The format names no functions either, so a Go function reports `coverage_pct` and `crap` as null, and its line counts as null unless a line-measuring artifact also covered the file |\n\nThe record shape and version behind each format are its row under \"Coverage artifacts\" in\n`${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fcollectors.md`, which carries each claim with its basis, date and\nrecheck trigger.\n\nThe SQLite data file coverage.py writes while measuring is never read; why is the coverage.py\nSQLite row of that file. Run `coverage json` or `coverage lcov` and point this skill at the result.\n\n## Run it\n\n```bash\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\"                                  # the change, with the artifact auto-discovered\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --artifacts build\u002Flcov.info src\u002F # an explicit artifact and an explicit scope\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --all --artifacts coverage.xml --artifacts cover.out\n\"${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh\" --json --all                     # the code-metrics\u002Fv2 document instead of markdown\n```\n\nWith no `--artifacts` and no `coverage.artifacts` in the config, the well-known names are looked\nfor under the repository root, at most two directory levels deep and never inside `node_modules`,\n`.git` or `vendor`: `coverage\u002Flcov.info`, `lcov.info`, `coverage.xml`, `cobertura.xml`,\n`coverage.json`, `coverage.out`, `cover.out`. Each artifact's format is detected from its content,\nnot its extension.\n\nPresent the markdown report as printed: the scope, the \"Coverage of this run\" table, the\nreferences with their provenance, then one row per file and per function. Keep the `--json`\ndocument when the numbers feed a comparison: pass it to `\u002Fverification:measure metrics` when the\n`verification` plugin is installed, treating `status: empty` on either side as INCONCLUSIVE;\notherwise keep the JSON beside your notes and compare by hand.\n\n## Getting a first artifact\n\nA run that finds no artifact reports every lane `unavailable`, lists the paths it searched, and\nends its markdown with a line pointing here, because the remedy is the project's to apply: this\nskill runs no test and installs nothing. One command per lane produces an artifact the next run\ncan read. Each row restates a producer's documented default, verified against the page in its\nBasis column on 2026-09-12; the producer's page wins over the row. Recheck trigger: a release note\nor changelog entry of that producer naming the flag or the output path in its row, or a read-time\nfetch of the Basis page that no longer states what the row does; either re-derives the row from the\npage and refreshes the date.\n\n| Lane | Producer | Command shape | Writes | Basis |\n|---|---|---|---|---|\n| TypeScript\u002FJavaScript | vitest | `npx vitest run --coverage --coverage.reporter=lcov` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered | [Vitest coverage guide](https:\u002F\u002Fvitest.dev\u002Fguide\u002Fcoverage.html), `coverage.reporter` |\n| TypeScript\u002FJavaScript | jest | `npx jest --coverage` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered; the default `coverageReporters` list carries `lcov` | [Jest CLI](https:\u002F\u002Fjestjs.io\u002Fdocs\u002Fcli), `--coverage`; [Jest configuration](https:\u002F\u002Fjestjs.io\u002Fdocs\u002Fconfiguration), `coverageReporters` default `[\"clover\", \"json\", \"lcov\", \"text\"]` |\n| TypeScript\u002FJavaScript | c8 or nyc, over any test runner | `npx c8 --reporter=lcov \u003Ctest command>` | lcov `.info` at `coverage\u002Flcov.info`, auto-discovered | [c8 README](https:\u002F\u002Fgithub.com\u002Fbcoe\u002Fc8), `--reporter`, which takes any [Istanbul reporter](https:\u002F\u002Fistanbul.js.org\u002Fdocs\u002Fadvanced\u002Falternative-reporters\u002F), `lcov` among them |\n| Python | coverage.py | `python -m coverage run -m pytest && python -m coverage json` | coverage.py JSON at `coverage.json`, auto-discovered; `coverage xml` writes `coverage.xml`, also auto-discovered; `coverage lcov` writes `coverage.lcov`, which needs `--artifacts` | coverage.py command pages [`run`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_run.html), [`json`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_json.html), [`xml`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_xml.html), [`lcov`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_lcov.html) |\n| Python | pytest-cov | `pytest --cov=\u003Cpackage> --cov-report=json` | coverage.py JSON at `coverage.json`, coverage.py's default name for the `json` report, auto-discovered; `--cov-report=json:\u003Cpath>` moves it | [pytest-cov reporting](https:\u002F\u002Fpytest-cov.readthedocs.io\u002Fen\u002Flatest\u002Freporting.html), `--cov-report`; the default file name is coverage.py's, per its [`json`](https:\u002F\u002Fcoverage.readthedocs.io\u002Fen\u002Flatest\u002Fcommands\u002Fcmd_json.html) page |\n| Bash | kcov | `kcov \u003Coutdir> bash \u003Ctest script>` | Cobertura-compatible XML under `\u003Coutdir>`; pass the `cobertura.xml` it writes with `--artifacts` | [kcov README](https:\u002F\u002Fgithub.com\u002FSimonKagstrom\u002Fkcov), \"Kcov will also write cobertura-compatible XML output\" |\n| Go | `go test` | `go test .\u002F... -coverprofile=coverage.out` | Go cover profile at the path `-coverprofile` names; `coverage.out` and `cover.out` are auto-discovered; file rows carry the statement ratio, function rows need a line artifact too | [`go` command, testing flags](https:\u002F\u002Fpkg.go.dev\u002Fcmd\u002Fgo#hdr-Testing_flags), `-coverprofile` |\n| C# | coverlet, as the `dotnet test` collector | `dotnet test --collect:\"XPlat Code Coverage\"` | Cobertura XML at `TestResults\u002F\u003Crun id>\u002Fcoverage.cobertura.xml`; pass it with `--artifacts`. The C# complexity lane is deferred, so the lane reports file rows and no CRAP | [coverlet VSTest integration](https:\u002F\u002Fgithub.com\u002Fcoverlet-coverage\u002Fcoverlet\u002Fblob\u002Fmaster\u002FDocumentation\u002FVSTestIntegration.md), `--collect:\"XPlat Code Coverage\"` |\n\nAuto-discovered means the output lands on one of the well-known names the run looks for with no\n`--artifacts`; anything else is named explicitly or listed under `coverage.artifacts` in the\nconfiguration.\n\n## Reading the numbers\n\n- A file row is the artifact's own line table: `coverage_pct`, `lines_executable`, `lines_hit`.\n  A file a Go cover profile covers is the exception: the profile weighs statements, so the\n  percentage is its statement ratio, `cov_source` reads `statement-ratio`, and both line counts\n  are null.\n- A function row exists for every function whose complexity collector reported a real end line.\n  It carries `coverage_pct`, `cyclomatic`, `crap`, and, in the JSON, `cov_source` and `hit`.\n- The markdown table is ordered so the rows worth reading survive its 200-row cap: rows over a\n  reference first, then function rows by CRAP descending with a null CRAP last, then file rows by\n  coverage ascending with a null percentage last, and file and line only as the tie-break. The\n  first function row is the highest-CRAP function in scope, however its file name sorts; the rows\n  the cap drops are counted on the table's last row and are all in the `--json` document, which is not\n  capped.\n- `cov_source` says where the coverage came from: `artifact-region` when the artifact carried the\n  function's own region (coverage.py `functions`, a Cobertura `\u003Cmethod>`, an lcov 2.2 `FNL` end\n  line), `line-range` when the range came from the complexity collector, in which case nested\n  function ranges are subtracted from the parent first, `statement-ratio` on a file row whose\n  percentage came from a Go profile's statements rather than from a line table, and `ambiguous`\n  on a function the join refused. The markdown table renders the values; the per-row `cov_source`\n  and `hit` fields are in the `--json` document.\n- A function the artifact names by a short name binds to the function whose line range holds the\n  lines the artifact recorded, not to whichever function the name happens to end at. Two `run`\n  methods in one file and a record named `run` are one function's coverage, and the range says\n  whose. Where nothing separates the candidates, because the artifact placed none of them at a\n  line or because two of them fall inside the same range, the join is refused rather than guessed:\n  the row reads `cov_source: ambiguous`; coverage, line counts, hit, and CRAP are null because\n  those numbers would be a guess, while cyclomatic from the complexity row is kept because that\n  collector already measured it; it carries a `coverage-ambiguous` label, and the lane's run row\n  turns `partial` and names the functions it left unjoined. A missing coverage number an operator\n  can see beats a wrong number they cannot.\n- `hit` is the artifact's function-hit flag. When it says the function was never entered,\n  `coverage_pct` is 0 rather than the 1\u002FN a declaration line executed at import would produce.\n- A function with no executable lines in the artifact reports `coverage_pct: null` and\n  `crap: null`, never 0: a fabricated 0 percent would produce a fabricated maximal CRAP. Its\n  `lines_executable` and `lines_hit` are 0 when an artifact that measures lines covers the file\n  and carries none in that range, and null when no artifact covering the file measures lines at\n  all, which is every function of a file only a Go profile covers.\n- CRAP is Savoia and Evans' formula (Agitar Labs, 2007). It is a description of two numbers you\n  already have, not a validated predictor of change risk, and no standard sets a threshold for it.\n  `\u002Fcode-metrics:principles` carries the provenance, the rename history, and the caveats; read it\n  before you quote a CRAP number at anyone.\n- A lane whose resolved complexity collector reports no function end lines gets a `\u003Clane>\u002Fcrap`\n  row with `status: not-applicable` and that reason. Bash is that lane in this version: no\n  maintained Bash collector reports an end line, so Bash CRAP is a stated gap, not a null.\n- A lane whose cyclomatic collector did not resolve, or ran and produced nothing parseable, gets a\n  `\u003Clane>\u002Fcrap` row with `status: unavailable` naming that collector and its reason, whatever the\n  lane's coverage row says. CRAP needs both numbers, and the missing one is the cause worth\n  reading; when the coverage row is not `ok` either, its reason follows the collector's on the\n  same row, so a partial count or a no-artifact search never hides a collector failure and the\n  collector never hides them. `Functions: 0` in the summary reads with that row.\n- A lane matched by fewer than all of its scope files carries `status: partial` and the reason\n  `partial, N of M scope files present in the artifacts`, so a total miss never reads as \"no\n  executable lines\" and the document cannot settle as `complete` while a row says `N of M`. A lane\n  no artifact covers is `unavailable` with that count; when no artifact was found at all, the\n  reason lists every path searched. A lane that left a function unjoined is `partial` for that\n  reason too.\n- A row carrying that `N of M` count also says which files: its reason names the first five scope\n  files no artifact mentions (`; missing: a, b, c, d, e, +N more in the JSON`) and the JSON row's\n  `missing` key lists all of them, root-relative and sorted. The count alone cannot tell a test\n  file an artifact never records from a source file the suites never reach; the paths can.\n- Whether the covered code is actually checked by its tests is a different question, and coverage\n  alone cannot answer it: a line can execute under a test that asserts nothing.\n  `\u002Fmutation-testing:audit` owns that question when the `mutation-testing` plugin is installed;\n  without it, say that the coverage number is an execution count and stop there.\n- Exit 0 whenever a report was produced, including an empty one; exit 2 for a usage error, which\n  includes a named artifact or scope path that does not exist; exit 3 when a complexity collector\n  ran and produced nothing parseable. On exit 3 the collector's stderr is on the `\u003Clane>\u002Fcrap`\n  row, and the summary carries an `Exit 3:` line naming the lane and collector, so the markdown\n  alone says why the run has no function rows.\n\n## Configuration\n\nEverything tunable resolves through the code-metrics layers (user-global, team\n`docs\u002Fconventions\u002Fcode-metrics.yaml`, personal override; per-key override; keys in `${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fconfig.md`):\n\n| Key | Default | Effect |\n|---|---|---|\n| `coverage.artifacts` | `[]` | Explicit artifact paths; empty means auto-discover. A listed path that does not exist is a usage error, so a stale entry is loud |\n| `coverage.path_prefix_strip` | `[]` | Prefixes removed from artifact paths before the join, for compiled-output layouts (`dist\u002F`, `build\u002F`) |\n| `coverage.reference` | `null` | No bar. A number counts rows below it as `over_reference` and nothing else |\n| `coverage.crap.reference` | `null` | No bar; no standard sets one for CRAP |\n\n`\u002Fcode-metrics:setup` writes the team file and probes the collectors.\n\n## What this skill does not do\n\n- It never runs tests, starts a test runner, or installs a coverage tool. It reads what a build\n  already wrote; if there is no artifact, the report says so and stops.\n- It never reads the SQLite data file coverage.py writes while measuring.\n- It never judges. A reference here is a value to count against, and no finding, severity, or\n  exit-code gate follows from one.\n- It does not measure branch, condition, or MC\u002FDC coverage. Those records are parsed past, not\n  reported, because the join is a line-level one.\n\n## Next\n\n- The numbers feed a before-and-after comparison: `\u002Fverification:measure metrics`.\n- Coverage looks healthy but the tests may assert nothing: `\u002Fmutation-testing:audit`.\n- A number is about to be quoted at someone: `\u002Fcode-metrics:principles`.\n\n## Gotchas\n\n- The complexity collector decides which functions can carry CRAP at all. With no collector\n  resolving for a lane, that lane has no function rows and its `\u003Clane>\u002Fcrap` row is `unavailable`\n  naming the collector's own reason. File-level coverage is unaffected: the scope the join keys on\n  is the one the audits measure, not the files a complexity collector emitted rows for, so a file\n  with no functions still reports its lines.\n- Path shapes differ between artifact and source tree. The join normalizes both sides, strips the\n  repository root and the configured prefixes, then falls back to a component-wise suffix match\n  and finally to a unique basename (which is how a Go profile's module path resolves). Both\n  fallbacks require a single winner: an artifact path that fits two scoped files equally well\n  (two services vendoring the same `pkg\u002Fa.py`) stays unmatched and shows up in the partial count,\n  because attributing it to either would credit one service with the other's coverage. The\n  basename fallback needs a single winner on the artifact side too, counted across every artifact\n  read rather than within one: when two services each ship a `handler.go` and only one is in\n  scope, the out-of-scope path stays unmatched instead of merging its coverage into the scoped\n  file, even though the scope names exactly one candidate.\n- An artifact older than the source it describes joins by line number and will be wrong without\n  saying so. Re-run the tests before reading the numbers when the diff has moved lines.\n- Two artifacts covering one file are merged by keeping the larger hit count per line, so a line\n  is never counted twice. Two artifacts covering one function are folded the same way: the hit\n  flag is the larger of the two, so a suite that never entered the function cannot report it at 0\n  percent beside a file row the other suite already showed as covered. Two Go profiles covering\n  one file merge block by block on the profile's own `start.col,end.col` identity and are totaled\n  after that, so two test shards that entered different blocks report what they covered between\n  them rather than what the luckier shard covered alone.\n- A Go profile and a line artifact naming the same `.go` file do not average or overwrite each\n  other: the profile's statement ratio is that file's exact native measure and wins the\n  percentage, the row says `statement-ratio`, and the line counts stay null rather than reading as\n  the counts behind a number they did not produce. The lane's `collector` still names both formats\n  read. Function rows are line-based, so a Go function does take its region from a line artifact\n  when one covers it.\n- A Cobertura report with several `\u003Csource>` roots is read against all of them: a relative class\n  filename takes the first root under which that path exists in the scanned tree, and falls back\n  to the first root when no candidate exists. Two roots that both hold the file is an ambiguity\n  the report cannot settle, so the first one wins and the class lands under it. The probe settles\n  anything only when the roots are relative, or absolute and present on the machine running the\n  audit. A report built elsewhere declares absolute roots (the usual coverlet, kcov and gcovr shape\n  for a CI run) naming directories this machine does not have, so every candidate misses and every\n  class lands under the first root. Rewriting a root from another machine onto the local tree\n  needs a mapping the report does not carry.\n- Two functions in one file whose qualified names share a tail (`A.run` and `B.run`) can bind to\n  each other's artifact region when the artifact carries only the short name; the file-level\n  numbers are unaffected.\n",{"shell":24,"metadata":25,"description":7,"allowed-tools":28,"argument-hint":31,"user-invocable":32,"disable-model-invocation":20},"bash",{"summary":26,"workflow-stage":27},"Coverage and CRAP read from build artifacts, no verdict","anytime",[29,30],"Bash(${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-coverage.sh:*)","Bash(git branch --show-current:*)","[--json] [--all] [--base \u003Cref>] [--artifacts \u003Cpath>]... [\u003Cpath>...]",true,"main","d0e68c3f36b92a1d6381f0ba0060d7a0b8b886c5","plugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage",7,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins","https:\u002F\u002Fgithub.com\u002Fmelodic-software","https:\u002F\u002Fmelodicsoftware.com","Independent software studio building developer tools, automation, and application services, with roots in music.","2026-11-15 05:15:06",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":32},[44,47,50],{"tool":45,"command":46},"npx","npx skills add melodic-software\u002Fclaude-code-plugins --skill audit-coverage",{"tool":48,"command":49},"bunx","bunx skills add melodic-software\u002Fclaude-code-plugins --skill audit-coverage",{"tool":51,"command":52},"pnpm","pnpm dlx skills add melodic-software\u002Fclaude-code-plugins --skill audit-coverage","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002FSKILL.md",[55,61,67,71,75,79,83],{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"evals\u002Fevals.json","REFERENCE","参考文档",2716,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fevals\u002Fevals.json",{"path":62,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":65,"sourceHtmlUrl":66},"scripts\u002Faudit-coverage.sh","SCRIPT","脚本",8144,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Faudit-coverage.sh",{"path":68,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":69,"sourceHtmlUrl":70},"scripts\u002Faudit-coverage.test.sh",16678,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Faudit-coverage.test.sh",{"path":72,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":73,"sourceHtmlUrl":74},"scripts\u002Fcrap.py",2451,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Fcrap.py",{"path":76,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":77,"sourceHtmlUrl":78},"scripts\u002Fjoin.py",48590,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Fjoin.py",{"path":80,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":81,"sourceHtmlUrl":82},"scripts\u002Ftest_crap.py",3486,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Ftest_crap.py",{"path":84,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":85,"sourceHtmlUrl":86},"scripts\u002Ftest_join.py",65188,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-metrics\u002Fskills\u002Faudit-coverage\u002Fscripts\u002Ftest_join.py",[88,95,102,109,114,121],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7DYYDG31B9YX7N9MB9QW2GW","track","Backlog reads through the bound tracker: search open and closed items, view one item with its comments and state history. Use when another skill searches or reads tracker items.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fevals\u002Frecall-still-reports-reopened-bug\u002Fwork-items-stub\u002Fskills\u002Ftrack\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7DYXVBAVG3NV30PA7GKW9Q6","reviewer-media","Reviewer media for one pull request: decides whether it needs a narrated walkthrough, builds the storyboard from \u002Freview:explain-change's digest, has \u002Fexplainer-video:produce render it and hands the manifest to \u002Fsource-control:pull-request media. Use when asked for 'reviewer media', 'a walkthrough video for this PR', 'a produced video of this change', or when pull-request ready or monitor hands a PR over.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Freview\u002Fskills\u002Freviewer-media\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7DYRZ5YD44AMYH26YR25GZ8","route","Model and effort per agent role (orchestrator, worker, verifier, retrieval) for a multi-agent run, resolved through the bundled defaults and the user, team and personal override config layers into fenced JSON with single-agent and fan-out variants and the layer behind each value. Use when asked 'which model for the subagents', 'route the roles' or 'model routing', or before a skill passes a role's model on a spawn.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-pin-without-override\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":110,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":113,"publishedAt":19},"01M7DYRTVT4FWTGDJY0HCCFJMZ","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-override-from-route\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":120,"publishedAt":19},"01M7DYNZC8PNW3T27V2ADKECKJ","principles","Module-design vocabulary for architecture findings, and the checks built on it. Use when: 'what is a deep module', 'what is a seam', 'deep vs shallow module', 'what is the deletion test', 'leverage and locality', 'module design vocabulary', 'which word do we use for X in architecture findings'. Scanning a codebase for friction is \u002Farchitecture:improve.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fskills\u002Fprinciples\u002FSKILL.md",[],[],{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":127,"publishedAt":19},"01M7DYNMCZNEWBFHV9KTSS6H60","write","Structured five-field bug report (title, steps to reproduce, expected vs actual, justified severity, suggested fix location) from an informal description, read-only. Use when the user names a defect they observed, or asks for one written up ('report a bug', 'file a bug', 'write this up as a bug').","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fevals\u002Fdeepening-defect-goes-to-bugs-write\u002Fbugs-stub\u002Fskills\u002Fwrite\u002FSKILL.md",[],[]]