agent-friendly-cli
Skill1 starsChanged 14 days agoMITView on GitHub
Design or review CLIs and companion skills for reliable agent use, including output contracts, authentication, transactions, distribution and self-update, and acceptance.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- briandconnelly/skillstheir own version, first seen Oct 4, 2026
- akhy/agent-skillstheir own version, first seen Mar 4, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- akhy/agent-skills★ 1same name
- briandconnelly/skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 25
Read the applicable design, implementation, or acceptance sections of implementation workflow, then the relevant references listed above.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent-Friendly CLI Development
Scope
This skill governs cross-language CLI contract design and acceptance: how to slice commands, shape output, rank exit codes, handle non-interactive use, avoid pitfalls, distribute and self-update, and run acceptance. Language-level implementation details (e.g. Go's cobra/flag, Python's argparse) belong to language-specific references (Go: references/go-engineering.md); the main line does not repeat them.
One-line goal: make the CLI low-token, low-ambiguity, low-risk, auditable, reproducible, and reversible for agents — and readable and interactive by default for humans. One CLI, two audiences — not two tools.
Which reference to read when
- Design the command surface / output contract / exit codes, or review whether a CLI is agent-friendly → read references/design-principles.md — the P0/P1/P2 tiers, dual-audience rules, the read-before-write fingerprint recipe, and contract-evolution red lines. When reviewing, check every item; each gap is a finding to report.
- Build it, and want the full workflow → read references/implementation-workflow.md — the P0 checklist, self-serve installation, ask-once login credentials, transaction skeleton, high-risk writes, and acceptance.
- Implement, and want to dodge real-world incidents → read references/pitfalls.md — 43 field-tested entries (dry-run must be truly read-only, idempotency signals, secret safety, test isolation, pagination traps, …), each as pitfall → consequence → rule.
- Distribute the CLI → read references/distribution.md — the minimal complete background self-update spec (forward-only version guard, throttling, kill switches), prebuilt-artifact and manifest discipline, version injection, the post-change closure loop, and companion-skill reverse sync.
- Work in Go → read references/go-engineering.md — the cross-compilation platform matrix, syscall platform split, ldflags version injection, and shared-package reuse rules.
- Finish building, ready for acceptance → read references/verification.md — the evidence-pasting acceptance checklist, real-agent evaluation methodology, and the post-launch session-history audit.
Selected workflow
Read the applicable design, implementation, or acceptance sections of implementation workflow, then the relevant references listed above. A design review does not start installation, login, or production writes. Keep transaction, authorization, and terminal-state boundaries for execution tasks.
Contract evolution
The JSON envelope, exit codes, and published field names are public contracts: after a version ships, only add, never change or remove. Breaking changes bump the interface version and get prominent callouts. Details at the end of design-principles.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.