feat: add skill-creator casedev skill#49
Conversation
The authoritative procedure for creating private org skills from a Linc chat: interview, draft (triggery summary, agent-directed content), approve, then create via 'casedev api POST /skills'. C3's Create with AI quick action points agents at this skill (skill_read skill-creator). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Oppenheimer Cleanup ReviewNo provable dead code, redundancy, or over-engineering in the new skill file; no safe simplifications identified without behavior risk. No cleanup opportunities found. This code is already lean. Cleanup review by Oppenheimer | Re-run with |
Judicial ReviewAdds a new skill definition that documents the workflow for creating private org skills via the casedev CLI. No security, data integrity, or architectural issues identified in the content as written. No significant issues found. This code is cleared to ship. Legal-Grade review by Thurgood | Re-review with |
DB Cooper Database ReviewDB review: no database schema, migration, or query changes in this PR. No DB risks detected. No database risks found. Migrations and DB-impacting changes look safe. DB review by DB Cooper | Re-run with |
1 participant
Adds
skills/casedev/skill-creator— the authoritative procedure agents follow when a user asks to create a reusable skill from chat (C3's Create with AI flow points at it viaskill_read skill-creator).Covers: interview → draft (pushy/triggery summary per skill-authoring best practices, agent-directed content) → explicit user approval → creation as a private org skill via
casedev api POST /skills(JSON-file pattern to avoid shell quoting) → plain-language confirmation. Includes update path (PUT /skills/<slug>), credential rules, and troubleshooting (slug conflict, auth failure, content cap).Verified the CLI surface:
casedev api <method> <path> --bodyis the raw-call form available in the pre-authenticated Linc sandbox (native skill tools are read-only).🤖 Generated with Claude Code