doc-auditlisted

## Scope for v1 - **Static analysis only.** Grep, file parsing, filesystem reads. Does not fetch URLs, does not run builds, does not execute code examples. - **No auto-fix.** Produces a markdown report; corrections are proposed but never applied. - **Live link check deferred.** HTTP/HTTPS targets are out of scope (flaky, rate-limited). Tracked as `DOC-live-links`. - **Semantic drift deferred.** Prose obsolescence, deprecation-age reasoning, and doc-bloat judgement are judgment-heavy and out of scope for v1. --- ## Configuration (adapt before first run) > Replace these placeholders: > - `[DOC_PATH]` - primary location of user-facing docs (e.g. `docs/`, `documentation/`, `website/`). Always included. > - `[README_PATH]` - repository README (usually `README.md` at project root). > - `[ADR_PATH]` - ADR directory if present (e.g. `docs/adr/`, `docs/decisions/`). Leave empty if none. --- ## Step 0 - Target resolution Parse `$ARGUMENTS` for a `target:` or `mode:` token. | Pattern | Meaning | |---|---| | `target:path:<dir>` | Audit docs under a specific directory only (e.g. `target:path:docs/api`) | | `target:file:<glob>` | Audit files matching the glob (e.g. `target:file:docs/**/*.md`) | | `mode:all` / no argument | **Full audit - README, CHANGELOG, every `.md` under `[DOC_PATH]`, and ADRs under `[ADR_PATH]` if set.** | **STRICT PARSING**: derive target ONLY from explicit text in `$ARGUMENTS`. Do NOT infer from conversation context, recent blocks, or memory. Announce: `Run