Install
$ agentstack add skill-scoobydont-666-shared-claude-skills-skill-updater ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Skill Updater — Self-Assessment & Auto-Update (Claude Code)
When to Trigger
- "update your skills" / "refresh skills" / "skill audit"
- "are your skills up to date" / "check your skills"
- "tune yourself up" / "make yourself better"
Workflow
Phase 1 — Inventory
- List all installed skills:
ls ~/.claude/skills/ - Read CLAUDE.md files:
~/.claude/CLAUDE.md, projectCLAUDE.md, nested - Scan recent git history:
git log --oneline -20for active workstreams
Phase 2 — Audit Each Skill
For every skill directory:
- Read the SKILL.md — check for:
- Stale dates or version references
- Overlap with other skills
- Missing coverage for known active projects
- Description quality (will it trigger reliably?)
- References to claude.ai-only tools (
Note in response,Grep/Read project memory,
Read session transcripts) — these don't exist in Claude Code
- Read reference files — check for:
- Outdated data (prices, versions, API endpoints, tax law)
- Missing reference files that should exist
- Freshness checks (web search only when needed):
- token-miser: Compare
references/model-pricing.mdheader date. If >90 days, web search. - cpa-tax-specialist: Check for major tax law changes if >12 months old.
- crypto-monero-wizard: Check monerod/P2Pool/XMRig versions if >6 months old.
- ai-cluster-knowledge: Compare documented hardware against current memory files.
Phase 3 — Gap Analysis
- Active projects without skills — check memory files and CLAUDE.md for gaps
- CLAUDE.md gaps — conventions in skills not reflected in CLAUDE.md
- Skill overlap — two skills covering the same ground
- Missing conventions — patterns in code/commits not captured anywhere
- Stale CLAUDE.md entries — rules referencing removed files or outdated versions
- Cross-config contradictions — invoke config-auditor skill for hardware specs, repo counts, port assignments, collection prefixes, project phases. SSH to verify hardware before changing specs.
Phase 4 — Report & Propose
Terse tables and bullet points:
- What's current — no changes needed
- What's stale — specific data points that are wrong
- What's missing — skills or entries that should exist
- What overlaps — skills to consolidate
- Proposed actions — ranked by impact
Execute non-breaking fixes immediately. Only ask about ambiguous items.
Phase 5 — Execute
- Skill updates — Edit SKILL.md and reference files in place
- New skills — Create directory + SKILL.md in
~/.claude/skills/ - Reference updates — Web search for current data, rewrite reference file
- CLAUDE.md updates — Edit the appropriate CLAUDE.md
- Sync to repo — Run
collect.shto update claude-config repo
Frugality
- Don't web-search every skill — only time-sensitive reference data
- Don't re-read skills already in the current session
- Only propose actions with clear ROI
- Batch file edits
Output Format
Tables and bullet points, not prose. Flag issues, propose fixes, execute on approval.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: scoobydont-666
- Source: scoobydont-666/shared-claude-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.