Skills Discovery Curation

Skills Discovery and Curation

Use this skill to build a practical skill portfolio for a repo.

Workflow

1. Route the request through $skill-hub for baseline chain selection.
2. Capture the current meta-harness authority contract for the curation pass:
   • PC Control evidence/status surfaces first
   • canonical repo root G:\GitHub on the maintainer workstation

• direct LM Studio authoritative model lane http://127.0.0.1:1234/v1
• http://127.0.0.1:2337/v1 only when explicitly selected; http://127.0.0.1:9000 is non-model MX3 support only

3. Inventory available skills and currently installed skills.
4. If maintaining multiple repos, run $skills-cross-repo-radar first.
5. When the ask is about recent Codex work, inspect recent local workstreams/continuity artifacts as evidence in addition to git history.
6. Map repository workflows to missing capabilities.
7. Include practical baseline checks via $skill-common-sense-engineering.
8. Run $usage-watcher and capture usage mode plus analysis/plan artifacts.
9. Run $skill-cost-credit-governor and capture analysis/policy artifacts.
10. Run $skill-cold-start-warm-path-optimizer and capture cold/warm analysis/plan artifacts.
11. Include quality/compliance lanes where needed ($skill-auditor, $skill-enforcer).
12. For candidate additions/updates, classify each skill as unique vs upgrade using $skill-auditor.
13. Require arbiter evidence status pass before admitting candidates.
14. Propose a minimal prioritized skill set (core first, optional second) with usage guardrail status.
15. Provide admission plan using local skill-arbiter lockdown flow.
16. For meta-harness or whole-PC curation, normalize repo paths to G:\GitHub, capture PC Control status/evidence first, and treat direct LM Studio http://127.0.0.1:1234/v1 as model authority, selecting http://127.0.0.1:2337/v1 explicitly and keeping :9000 support-only.

Discovery Commands

find $CODEX_HOME/skills -mindepth 1 -maxdepth 1 -type d -printf '%f\n' | sort
find $CODEX_HOME/skills -mindepth 1 -maxdepth 2 -name SKILL.md -type f | sort

Optional recurring cross-repo scan:

python3 "$CODEX_HOME/skills/skills-cross-repo-radar/scripts/repo_change_radar.py" \
  --repo "<PRIVATE_REPO_A>=/path/to/<PRIVATE_REPO_A>" \
  --repo "<PRIVATE_REPO_B>=/path/to/<PRIVATE_REPO_B>" \
  --since-days 14 \
  --json-out /tmp/skills-cross-repo-radar.json

Curation Output Format

1. Current inventory summary.
2. Missing capability gaps by repo workflow.
3. Usage guardrail decision (economy/standard/surge) with JSON evidence paths from usage-watcher, skill-cost-credit-governor, and skill-cold-start-warm-path-optimizer.
4. Recommended skill candidates (ranked).
5. Admission command using --personal-lockdown.

Admission Command Template

python3 "$CODEX_HOME/skills/skill-arbiter/scripts/arbitrate_skills.py" <skill> [<skill> ...] \
  --source-dir "$CODEX_HOME/skills" \
  --window 10 --threshold 3 --max-rg 3 \
  --personal-lockdown

Scope Boundary

Use this skill only for the skills-discovery-curation lane and workflow defined in this file and its references.

Do not use this skill for unrelated lanes; route those through $skill-hub and the most specific matching skill.

Reference

• references/discovery-checklist.md

Loopback

If this lane is unresolved, blocked, or ambiguous:

1. Capture current evidence and failure context.
2. Route back through $skill-hub for chain recalculation.
3. Resume only after the updated chain returns a deterministic next step.