agent-clilisted

Use when building, designing, reviewing, or evaluating command-line tools for AI agent consumption -- structures commands with machine-readable JSON output on stdout, human diagnostics on stderr, adds --json/--fields/--dry-run/--quiet flags, implements semantic exit codes, enforces TTY-aware dual-mode output, hardens input against path traversal and injection, creates CONTEXT.md and AGENTS.md agent knowledge files, adds --help-json schema introspection, and scores CLI agent-friendliness on a 0-21 rubric across 7 axes. Triggers on: 'agent-friendly CLI', 'CLI for agents', 'build a CLI', 'design CLI commands', 'machine-readable output', '--help-json', 'CLI scoring', 'agent DX', 'context window discipline', 'evaluate CLI', 'CLI agent readiness', 'score this CLI'. NOT for TUI/full-screen apps. NOT for GUI apps. NOT for REST/GraphQL API design.
wagneripjr/skills · ★ 0 · AI & Automation · score 75

Install: claude install-skill wagneripjr/skills

# Agent-Friendly CLI Design & Evaluation Build command-line tools that AI agents consume reliably and humans use comfortably. These goals are orthogonal — a CLI can serve both audiences from the same command surface by detecting context (TTY vs pipe) and adapting output. **Build mode** (default) — walk through Phases 1-6 to construct an agent-first CLI from scratch or retrofit an existing one. **Evaluate mode** — activated when the user says "score", "evaluate", "audit", or "rate" — jump to Phase 7 to score a CLI on the 7-axis rubric. This skill targets **command-based CLIs** (like `git`, `docker`, `gh`, `kubectl`) — tools with subcommands, flags, and structured output. Not for full-screen TUI applications, dashboards, or GUI tools. ## Quick Decision Guide | Need | Approach | Reference | |------|----------|-----------| | Design command grammar and hierarchy | Noun-verb or verb-noun pattern, 2-3 levels max | [command-design.md](references/command-design.md) | | Add `--json` flag with consistent envelope | `{ "status", "data", "error", "meta" }` on every command | [output-design.md](references/output-design.md) | | Stream results without buffering | NDJSON — one JSON object per `\n`-separated line | [output-design.md](references/output-design.md) | | Reduce agent token consumption | `--fields`, `--quiet`, `--limit`, `--summary` flags | [output-design.md](references/output-design.md) | | Accept raw JSON payloads | `--data '{"key":"val"}'` or stdin pipe alongside flags | [in