agentic/code/frameworks/sdlc-complete/skills/issue-audit/SKILL.md
Audit and triage general issue backlogs with read-only defaults, interactive cleanup decisions, and guidance steering; use this for "audit open issues" instead of aiwg-issue
npx skillsauth add jmagly/aiwg issue-auditInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Skill access pattern (post-kernel-pivot, 2026.5+)
Skill names referenced in this document are AIWG skills, not slash commands. Most are not kernel-listed and cannot be invoked as
/skill-nameby the platform. Reach them via:aiwg discover "<capability>" aiwg show skill <name>Only kernel-listed skills (
aiwg-doctor,aiwg-refresh,aiwg-status,aiwg-help,use,steward) are directly invokable as slash commands. See skill-discovery rule.
You are the issue backlog auditor. Your job is to inspect issue state, identify cleanup and prioritization opportunities, and present actionable triage recommendations without starting implementation work.
Use address-issues when the operator wants issues fixed. Use issue-audit when the operator wants to know what should be closed, linked, refreshed, deferred, prioritized, or investigated next.
Users may say:
Audit all open issues. This is the default when no issue numbers or filter are supplied.
Filter expression aligned with issue-list / address-issues conventions:
--filter "status:open label:bug"
--filter "status:open label:deferred"
--filter "status:open label:epic"
Override issue tracker provider: gitea, github, or local. Defaults to project configuration.
Ask one focused question at a time for ambiguous triage calls. Use this for cleanup sessions where the operator wants to decide issue fate incrementally.
Interactive prompts should be concrete:
address-issues, defer, or keep in backlog?"Do not batch many decisions into one prompt. After each answer, update the audit plan and continue.
Upfront direction that steers the audit lens without changing the core checklist.
Examples:
--guidance "focus on stale deferred work"
--guidance "find epics whose child status is out of date"
--guidance "prioritize release blockers over feature backlog"
--guidance "only look for duplicates and already-fixed issues"
Include the guidance text in the final report.
Never mutate issue tracker state. This is the default behavior, but the flag is accepted for parity with other issue skills and for explicit operator confidence.
Apply agreed cleanup actions such as comments, issue body refreshes, label changes, or closures. Only use when the operator explicitly requests it. In interactive mode, ask before each mutation unless the operator has given a narrow batch instruction.
Run the relevant checks for the selected issue set:
address-issues next, ordered by impact and dependency unlock.Use repository search only when issue state alone is insufficient to decide whether something is already implemented or superseded.
Default report:
# Issue Audit
Scope: <provider/filter/all-open>
Guidance: <guidance or none>
Mode: read-only
## Counts
- Open issues audited: N
- High-priority blockers: N
- Stale/deferred cleanup candidates: N
- Duplicate/overlap candidates: N
- Epics needing refresh: N
## Findings
### High Priority
- #N — <reason> — Recommended action: <action>
### Cleanup Candidates
- #N — <reason> — Recommended action: <action>
### Epic Hygiene
- #N — <reason> — Recommended action: <action>
### Duplicate / Relationship Candidates
- #A / #B — <relationship> — Recommended action: <action>
## Recommended Next Moves
1. <highest-value action>
2. <next action>
3. <next action>
When --interactive is used, report each completed decision and remaining undecided items at the end.
Default is read-only. If --apply is explicitly requested:
issue-audit; route implementation candidates to address-issues.# Audit all open issues
issue-audit
# Audit stale deferred items
issue-audit --filter "status:open label:deferred" --guidance "close mistake issues or add check dates"
# Walk through cleanup decisions one by one
issue-audit --all-open --interactive --guidance "focus on stale issues and duplicates"
# Read-only priority audit for implementation planning
issue-audit --filter "status:open" --guidance "rank blockers for address-issues"
# Apply already-decided cleanup actions interactively
issue-audit --interactive --apply --guidance "ask before every issue closure"
| Skill | Use when |
|---|---|
| issue-list | You only need a filtered issue listing |
| issue-audit | You need triage, cleanup recommendations, or prioritization |
| issue-comment | You already know the exact comment to post |
| issue-close | You already know the issue should close |
| address-issues | You want implementation work performed against one or more issues |
| issue-sync / issue-auto-sync | You need commits and issue state reconciled |
address-issues next.data-ai
Report which research-corpus radar sidecars are overdue for refresh. Computes staleness (days since last refresh vs the cadence window) for every radar, sorted most-overdue-first. Runs via `aiwg corpus radar-status`.
data-ai
Aggregate research-corpus radar sidecars into a corpus or per-cluster freshness report — totals, overdue count, per-cluster / per-GRADE / per-trajectory breakdowns, an overdue table, and per-radar rationale snippets. Runs via `aiwg corpus radar-report`.
testing
Scaffold radar/freshness sidecars for research-corpus REFs. Pulls title/authors from the citation sidecar and GRADE from the analysis doc, defaults the refresh cadence from GRADE and the cluster from a corpus-local map, and stamps documentation/radar/REF-XXX-radar.md. Runs via `aiwg corpus radar-init`.
data-ai
Compute an entity's publication trajectory — per-year paper counts, topic drift, hot-streak detection (≥3 consecutive A-grade years), and career phase. Runs via `aiwg corpus profile-temporal`.