From carta-cc
Audit repository documentation for structural and semantic issues, assign stable AUDIT-NNN IDs, and update docs/AUDIT_REPORT.md and docs/BACKLOG/TRIAGE.md.
How this skill is triggered — by the user, by Claude, or both
Slash command
/carta-cc:doc-auditThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Audit repository documentation for structural and semantic issues, assign stable AUDIT-NNN IDs, and update `docs/AUDIT_REPORT.md` and `docs/BACKLOG/TRIAGE.md`.
Audit repository documentation for structural and semantic issues, assign stable AUDIT-NNN IDs, and update docs/AUDIT_REPORT.md and docs/BACKLOG/TRIAGE.md.
Which command? This skill (
/doc-audit) audits documentation structure + semantics. For embedded-data integrity usecarta audit/carta doctor; for retrieval quality usecarta eval. See the "Which audit command?" table in README.
Check if .carta/scan-results.json exists and was written less than 1 hour ago (compare run_at timestamp to current time).
carta scan
Uses the installed
cartaCLI. If it isn't on your PATH, runpython -m carta scaninstead (from the repo root).
Read scan-results.json and extract:
issues list (structural findings)changed_since_last_audit list (docs to semantically check)statsCheck if docs/AUDIT_REPORT.md exists.
<!-- audit_counter: N --> value (increment by 1 for this run).For each entry in issues from the scan results:
new (no prior entry), persisting (was active last audit), resolved (was active, no longer in issues), needs-input (ambiguous — flag for human review).For each doc in changed_since_last_audit:
If Qdrant is reachable (run carta search "test" 2>&1 | head -1 — no error means reachable):
If Qdrant is unreachable, skip this agent and note "Qdrant agent skipped — collection unreachable" in the report.
resolved.Write docs/AUDIT_REPORT.md (create docs/ if missing) with this structure:
# Doc Audit Report
<!-- audit_counter: <N> -->
<!-- Last run: <YYYY-MM-DD> | Audit #<N> | Issues: <X> active, <Y> resolved, <Z> archived -->
## Active Issues
[active + persisting + needs-input issues, newest first]
## Resolved (this audit)
[issues resolved this run]
## Archive
<!-- Issues resolved for 2+ audits. Kept for history. -->
[archived issues]
Each issue block format:
### AUDIT-NNN <emoji> <status> — <context>
**Type:** <type>
**Doc:** `path/to/doc.md` (or **Docs:** for conflicts)
**Detail:** <detail text>
**Action:** <what to do>
**Backlog:** [DOC-NNN](BACKLOG/TRIAGE.md#DOC-NNN) ← only if linked
Emoji key: 🆕 new | ⚠️ persisting | 🔵 needs-input | ✅ resolved
For each issue that is new or newly needs-input AND does not already have a **Backlog:** entry:
docs/BACKLOG/TRIAGE.md, find the highest ### DOC-(\d+) number. Start from DOC-001 if none.### DOC-NNN [doc-audit] <short description>
**Source:** AUDIT-NNN (<status> since audit #N)
**Type:** <type>
**Docs:** `doc` → `related_doc` (or just `doc` for non-conflict issues)
**Action:** <specific action for a developer to take>
**Backlog:** [DOC-NNN](BACKLOG/TRIAGE.md#DOC-NNN) to the matching issue block in docs/AUDIT_REPORT.md.Issues flagged needs-input get a [needs-input] note appended after the title.
Report a summary:
"Audit #N complete. X new issues, Y persisting, Z resolved. docs/AUDIT_REPORT.md updated. N items appended to docs/BACKLOG/TRIAGE.md."
npx claudepluginhub ian-q/carta --plugin carta-ccFetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Applies a firm's KYC/AML rules grid to parsed onboarding records: assigns risk rating, checks required documents, outputs rule outcomes with citations, and routes for escalation.
Generates daily or weekly digests of activity from connected sources (chat, email, docs, tasks, CRM), highlighting action items, decisions, mentions, and project updates.