agent-friendly-cli-skill
AGENTS.md1 starsChanged 14 days agoMITView on GitHub
This repository is the source and release repo of the agent-friendly-cli skill. Follow these conventions when working in it.
In this repository2 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-friendly-cliSkill
The file
Files
AGENTS.md
This repository is the source and release repo of the agent-friendly-cli skill. Follow these conventions when working in it.
Repository structure
- The repo root is the skill body (
SKILL.mdrouter +references/six references +agents/openai.yaml+tests/regression fixtures). The local development authority is~/.config/agentsync/skills/agent-friendly-cli/— land changes in the authority first, then sync to this repo; never edit only the repo copy.
Hard constraints
- English only (user ruling, 2026-10-29): every file in this repo — skill body, references, tests, commit messages — must be English.
README.mdis the single exception and may stay bilingual. Never introduce Chinese anywhere else, including "temporary" notes. - Merge before push, never blind-push: before every push, clone the remote and
diff -rqit against the local authority — the remote may hold updates pushed from other machines (including remote-only fixes such as README install URLs). Absorb remote-only content into the local authority first, then sync-push; direct rsync overwrites are forbidden. - Commit identity, both fields, is x0c (
GIT_COMMITTER_NAME/GIT_COMMITTER_EMAIL+--author, x0c@users.noreply.github.com); real names and emails never enter commit history. - No internal-network content: class names, paths, and configs of internal platforms must not enter this repo; use generic phrasing for examples (no company or platform names), and generic references ("a scheduling-platform CLI") with dates kept in incident narratives.
Documentation map
README.md: the storefront — the skill's use cases and installation for outsiders. Keep the clone URL in sync on repo renames or install changes; never lose the Quick Install section.SKILL.md: the router (scope, which-reference-when, contract evolution) — keep it lean.references/design-principles.md: the design standard (P0/P1/P2 + dual audience + read-before-write fingerprint recipe).references/implementation-workflow.md: the workflow.references/pitfalls.md: field incidents.references/distribution.md: distribution and lifecycle.references/go-engineering.md: Go engineering.references/verification.md: acceptance.tests/: the skill's own regression fixtures (a fresh-machine simulation brief plus the pre-registered acceptance criteria). Re-run them per their README after changing the skill.
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.

