claude-skills / delegate
CLAUDE.md0 starsChanged 30 days agoMITView on GitHub
Rules bound to /delegate's own pipeline. Cross-cutting rules (background subagents, model choice) live in the root CLAUDE.md.
In this repository13 other files it gives its agents.
- audit/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- plan-it/CLAUDE.mdCLAUDE.md
- screens/CLAUDE.mdCLAUDE.md
- ship/CLAUDE.mdCLAUDE.md
- store-assets/CLAUDE.mdCLAUDE.md
- auditSkill
- delegateSkill
- plan-itSkill
- readinessSkill
- screensSkill
- shipSkill
- store-assetsSkill
The file
Files
/delegate internals
Rules bound to /delegate's own pipeline. Cross-cutting rules (background subagents, model
choice) live in the root CLAUDE.md.
Gotchas
/delegateomitsmodel:on purpose. It inherits the session model so analysis + review run on the strongest model available (currently Opus 5.5); an explicitmodel:pin would fix the skill to that model instead of tracking whatever the session runs on. The executor is always dispatched as sonnet; the orchestrator has no Edit/Write in allowed-tools (advisor-only, enforced by tooling). Large/architectural tasks get an AskUserQuestion gate offering /plan-it first.- test-writer owns the oracle, the executor never edits locked files (Phase 3.5, 2026-10-08). For a
[logic test]step the tests come fromtest-writerin oracle mode, written from a contract and red before the executor starts. Their files plustests/Pest.php,tests/TestCase.php,phpunit.xml(.dist)are hash-locked (audit/bin/oracle-lock.sh); Phase 5 fails onCHANGED, and onNONEwhileORACLE_EXPECTED=1. Worktree mode (--worktree) has no oracle in v1: a worktree only holds committed files. Detail:references/oracle.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.