Install
$ agentstack add skill-mmusa-second-nature-curate ✓ 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
Curate — keep the skill library sharp
Audit every skill the user has, score it, and propose keep / update / merge / retire actions. Present the report first; apply only what the user approves. Nothing is ever deleted — retired skills are archived.
Steps
1. Inventory
Collect from three places:
- User skills:
~/.claude/skills/*/SKILL.md - Project skills:
/.claude/skills/*/SKILL.mdfor the current project - Pending proposals:
~/.claude/second-nature/proposals/*/
Skip skills that belong to installed plugins/marketplaces (they update upstream; flag them only if broken). Identify plugin-owned skills by their path — only audit skills that live directly in the directories above.
2. Score each skill
For each, assess:
- Usage — is it ever invoked? Search recent transcripts (
~/.claude/projects/*/*.jsonl, last 60 days) for its name. Extract invocations cheaply with grep; do not read transcripts whole. - Staleness — do file paths, commands, tools, and URLs it references still exist? Check the load-bearing ones.
- Overlap — does another skill cover the same ground? Candidates to merge.
- Quality — vague descriptions (won't trigger reliably), missing trigger phrases, instructions that have drifted from how the user actually works now.
For pending proposals: flag any older than 30 days as promote-or-drop decisions.
3. Report, then act on approval
Present one table: skill, verdict (keep / update / merge / retire / promote-or-drop), and a one-line reason grounded in the evidence above. Recommendations only — then ask the user what to apply (they can approve all, some, or none).
For approved actions:
- Retire — move the skill directory to
~/.claude/second-nature/retired/-/. Neverrm. - Merge — write the combined skill, retire the absorbed one (same archive rule).
- Update — edit the SKILL.md in place; keep the user's voice and format.
- Drop proposal — move it to
retired/the same way.
Append every applied action to ~/.claude/second-nature/LOG.md (YYYY-MM-DD curate: retired — ).
4. Close
Summarize what changed and what was left alone. If the library is healthy, say so — a no-op curate is a valid result. Suggest (once, not naggingly) scheduling /harvest + /curate weekly if the user hasn't.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: mmusa
- Source: mmusa/second-nature
- 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.