claude-skills / ship
CLAUDE.md0 starsChanged 30 days agoMITView on GitHub
Rules bound to /ship's own pipeline. Cross-cutting rules (marker hashing, test-command:/ deploy-command: trust boundary) live in the root CLAUDE.md.
- Reads credentials
In this repository13 other files it gives its agents.
- audit/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- delegate/CLAUDE.mdCLAUDE.md
- plan-it/CLAUDE.mdCLAUDE.md
- screens/CLAUDE.mdCLAUDE.md
- store-assets/CLAUDE.mdCLAUDE.md
- auditSkill
- delegateSkill
- plan-itSkill
- readinessSkill
- screensSkill
- shipSkill
- store-assetsSkill
The file
Files
/ship internals
Rules bound to /ship's own pipeline. Cross-cutting rules (marker hashing, test-command:/
deploy-command: trust boundary) live in the root CLAUDE.md.
Gotchas
/ship's docs phase runs before the commit on purpose. Phase 0.8 inship/SKILL.mdinventories the repo's own docs (README, CLAUDE.md, CHANGELOG,docs/**, help pages,.env.example,SKILL.md), runscheck-docs-claims.shfor the mechanical part, and updates whatever the diff invalidated. It sits before Phase 1 so the change and its documentation land in one commit rather than in a follow-up nobody writes; Phase 1'sgit add -upicks the edits up, since every file it may touch is already tracked. It never creates a doc file that does not exist (this repo has no CHANGELOG, and/shipmust not invent one) and it never rewrites prose the diff did not invalidate. Since the 2026-10-03/auditrebuild there is no audit docs worker any more, so this phase is the only place docs get synced. When it edits a doc, onefact-checkeragent checks the changed claims against the code before the commit (added 2026-10-03; no doc edits, no dispatch)./shiptherefore needsEditinallowed-tools; it previously had onlyRead/Write, and a whole-fileWriteis the wrong instrument for a surgical README correction. Doc-only edits deliberately do not invalidate a fresh audit marker: they are prose by the same definitionclassify-diff.shuses, so Phase 2 keeps checking marker age only.
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.

