auditlisted

# docaudit:audit — change-driven documentation audit Report-only orchestrator. Reads the per-project adapter `${CLAUDE_PROJECT_DIR}/.claude/doc-audit.json`. If that file is absent, tell the user to run `/docaudit:init` (Plan 3) or that this repo has no adapter yet — do NOT invent project facts. `SD="${CLAUDE_SKILL_DIR}"` ; `CFG="${CLAUDE_PROJECT_DIR}/.claude/doc-audit.json"`. Also bind `ANCHOR_PATH="$(python3 -c 'import json,sys; print(json.load(open(sys.argv[1])).get("anchorPath",""))' "$CFG")"` for Phase 5. `--full` forces whole-corpus mode (ignores the anchor diff scope). ## Phase 0 — index preflight (deterministic) Run: `bash "$SD/scripts/mdq-index.sh" --config "$CFG" --repo-root "$CLAUDE_PROJECT_DIR"`. Parse `{mdqAvailable, reason, bin}` and bind `MDQ_AVAILABLE` (true/false) for Phase 3 and `MDQ_BIN` (the `bin` field, default `mdq`). `mdqAvailable:false` is EXPECTED, not an error (`reason` is `not-installed` / `disabled-by-config` / `index-failed`): record the reason and proceed in grep-degrade mode — the engine is fully functional without mdq. When `mdqAvailable:true`, the whole repo's Markdown is now indexed under `$CLAUDE_PROJECT_DIR/.mdq/index.sqlite`; indexing runs in a subprocess, so doc bodies never enter context — only this JSON summary does. This phase always runs first (both incremental and `--full`). When `MDQ_AVAILABLE` is true, also run `python3 "$SD/scripts/mdq-health.py" --bin "<MDQ_BIN>" --db "$CLAUDE_PROJECT_DIR/.mdq/index.sqlite"` and bind `MDQ_HEAL