[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BYCH9CX7AY4PBVT7R1T7AF":3,"skill-same:01M7BYCH9CX7AY4PBVT7R1T7AF":50},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":32,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":34,"updatedAt":35,"install":36,"resources":49},"01M7BYCH9CX7AY4PBVT7R1T7AF","ai-native-cli",null,"Design spec with 98 rules for building CLI tools that AI agents can safely use. Covers structured JSON output, error handling, input contracts, safety guardrails, exit codes, and agent self-description.","YYC-Cube\u002FYYC3-Knowledge-Base","YYC-Cube","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F235006543?v=4","01M79J2PH4SWD2X0825X6PN748","yyc3","01M79J2PHVK3F6ZBYT8K2SCZCW",1,"MIT","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-Skills-技能包\u002FYYC3-Skills-技能包-antigravity-awesome-skills\u002Fplugins\u002Fantigravity-awesome-skills-claude\u002Fskills\u002Fai-native-cli\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ai-native-cli\ndescription: \"Design spec with 98 rules for building CLI tools that AI agents can safely use. Covers structured JSON output, error handling, input contracts, safety guardrails, exit codes, and agent self-description.\"\nrisk: safe\nsource: https:\u002F\u002Fgithub.com\u002FChaosRealmsAI\u002Fagent-cli-spec\ndate_added: \"2026-03-15\"\n---\n\n# Agent-Friendly CLI Spec v0.1\n\nWhen building or modifying CLI tools, follow these rules to make them safe and\nreliable for AI agents to use.\n\n## Overview\n\nA comprehensive design specification for building AI-native CLI tools. It defines\n98 rules across three certification levels (Agent-Friendly, Agent-Ready, Agent-Native)\nwith prioritized requirements (P0\u002FP1\u002FP2). The spec covers structured JSON output,\nerror handling, input contracts, safety guardrails, exit codes, self-description,\nand a feedback loop via a built-in issue system.\n\n## When to Use This Skill\n\n- Use when building a new CLI tool that AI agents will invoke\n- Use when retrofitting an existing CLI to be agent-friendly\n- Use when designing command-line interfaces for automation pipelines\n- Use when auditing a CLI tool's compliance with agent-safety standards\n\n## Core Philosophy\n\n1. **Agent-first** -- default output is JSON; human-friendly is opt-in via `--human`\n2. **Agent is untrusted** -- validate all input at the same level as a public API\n3. **Fail-Closed** -- when validation logic itself errors, deny by default\n4. **Verifiable** -- every rule is written so it can be automatically checked\n\n## Layer Model\n\nThis spec uses two orthogonal axes:\n\n- **Layer** answers rollout scope: `core`, `recommended`, `ecosystem`\n- **Priority** answers severity: `P0`, `P1`, `P2`\n\nUse layers for migration and certification:\n\n- **core** -- execution contract: JSON, errors, exit codes, stdout\u002Fstderr, safety\n- **recommended** -- better machine UX: self-description, explicit modes, richer schemas\n- **ecosystem** -- agent-native integration: `agent\u002F`, `skills`, `issue`, inline context\n\nCertification maps to layers:\n\n- **Agent-Friendly** -- all `core` rules pass\n- **Agent-Ready** -- all `core` + `recommended` rules pass\n- **Agent-Native** -- all layers pass\n\n## How It Works\n\n### Step 1: Output Mode\n\nDefault is agent mode (JSON). Explicit flags to switch:\n\n```bash\n$ mycli list              # default = JSON output (agent mode)\n$ mycli list --human      # human-friendly: colored, tables, formatted\n$ mycli list --agent      # explicit agent mode (override config if needed)\n```\n\n- **Default (no flag)** -- JSON to stdout. Agent never needs to add a flag.\n- **--human** -- human-friendly format (colors, tables, progress bars)\n- **--agent** -- explicit JSON mode (useful when env\u002Fconfig overrides default)\n\n### Step 2: agent\u002F Directory Convention\n\nEvery CLI tool MUST have an `agent\u002F` directory at its project root. This is the\ntool's identity and behavior contract for AI agents.\n\n```\nagent\u002F\n  brief.md          # One paragraph: who am I, what can I do\n  rules\u002F            # Behavior constraints (auto-registered)\n    trigger.md      # When should an agent use this tool\n    workflow.md     # Step-by-step usage flow\n    writeback.md    # How to write feedback back\n  skills\u002F           # Extended capabilities (auto-registered)\n    getting-started.md\n```\n\n### Step 3: Four Levels of Self-Description\n\n1. **--brief** (business card, injected into agent config)\n2. **Every Command Response** (always-on context: data + rules + skills + issue)\n3. **--help** (full self-description: brief + commands + rules + skills + issue)\n4. **skills \\\u003Cname\\>** (on-demand deep dive into a specific skill)\n\n## Certification Requirements\n\nEach level includes all rules from the previous level.\nPriority tag `[P0]`=agent breaks without it, `[P1]`=agent works but poorly, `[P2]`=nice to have.\n\n### Level 1: Agent-Friendly (core -- 20 rules)\n\nGoal: CLI is a stable, callable API. Agent can invoke, parse, and handle errors.\n\n**Output** -- default is JSON, stable schema\n- `[P0]` O1: Default output is JSON. No `--json` flag needed\n- `[P0]` O2: JSON MUST pass `jq .` validation\n- `[P0]` O3: JSON schema MUST NOT change within same version\n\n**Error** -- structured, to stderr, never interactive\n- `[P0]` E1: Errors -> `{\"error\":true, \"code\":\"...\", \"message\":\"...\", \"suggestion\":\"...\"}` to stderr\n- `[P0]` E4: Error has machine-readable `code` (e.g. `MISSING_REQUIRED`)\n- `[P0]` E5: Error has human-readable `message`\n- `[P0]` E7: On error, NEVER enter interactive mode -- exit immediately\n- `[P0]` E8: Error codes are API contracts -- MUST NOT rename across versions\n\n**Exit Code** -- predictable failure signals\n- `[P0]` X3: Parameter\u002Fusage errors MUST exit 2\n- `[P0]` X9: Failures MUST exit non-zero -- never exit 0 then report error in stdout\n\n**Composability** -- clean pipe semantics\n- `[P0]` C1: stdout is for data ONLY\n- `[P0]` C2: logs, progress, warnings go to stderr ONLY\n\n**Input** -- fail fast on bad input\n- `[P1]` I4: Missing required param -> structured error, never interactive prompt\n- `[P1]` I5: Type mismatch -> exit 2 + structured error\n\n**Safety** -- protect against agent mistakes\n- `[P1]` S1: Destructive ops require `--yes` confirmation\n- `[P1]` S4: Reject `..\u002F..\u002F` path traversal, control chars\n\n**Guardrails** -- runtime input protection\n- `[P1]` G1: Unknown flags rejected with exit 2\n- `[P1]` G2: Detect API key \u002F token patterns in args, reject execution\n- `[P1]` G3: Reject sensitive file paths (*.env, *.key, *.pem)\n- `[P1]` G8: Reject shell metacharacters in arguments (; | && $())\n\n### Level 2: Agent-Ready (+ recommended -- 59 rules)\n\nGoal: CLI is self-describing, well-named, and pipe-friendly. Agent discovers capabilities and chains commands without trial and error.\n\n**Self-Description** -- agent discovers what CLI can do\n- `[P1]` D1: `--help` outputs structured JSON with `commands[]`\n- `[P1]` D3: Schema has required fields (help, commands)\n- `[P1]` D4: All parameters have type declarations\n- `[P1]` D7: Parameters annotated as required\u002Foptional\n- `[P1]` D9: Every command has a description\n- `[P1]` D11: `--help` outputs JSON with help, rules, skills, commands\n- `[P1]` D15: `--brief` outputs `agent\u002Fbrief.md` content\n- `[P1]` D16: Default JSON (agent mode), `--human` for human-friendly\n- `[P2]` D2\u002FD5\u002FD6\u002FD8\u002FD10: per-command help, enums, defaults, output schema, version\n\n**Input** -- unambiguous calling convention\n- `[P1]` I1: All flags use `--long-name` format\n- `[P1]` I2: No positional argument ambiguity\n- `[P2]` I3\u002FI6\u002FI7: --json-input, boolean --no-X, array params\n\n**Error**\n- `[P1]` E6: Error includes `suggestion` field\n- `[P2]` E2\u002FE3: errors to stderr, error JSON valid\n\n**Safety**\n- `[P1]` S8: `--sanitize` flag for external input\n- `[P2]` S2\u002FS3\u002FS5\u002FS6\u002FS7: default deny, --dry-run, no auto-update, destructive marking\n\n**Exit Code**\n- `[P1]` X1: 0 = success\n- `[P2]` X2\u002FX4-X8: 1=general, 10=auth, 11=permission, 20=not-found, 30=conflict\n\n**Composability**\n- `[P1]` C6: No interactive prompts in pipe mode\n- `[P2]` C3\u002FC4\u002FC5\u002FC7: pipe-friendly, --quiet, pipe chain, idempotency\n\n**Naming** -- predictable flag conventions\n- `[P1]` N4: Reserved flags (--agent, --human, --brief, --help, --version, --yes, --dry-run, --quiet, --fields)\n- `[P2]` N1\u002FN2\u002FN3\u002FN5\u002FN6: consistent naming, kebab-case, max 3 levels, --version semver\n\n**Guardrails**\n- `[P1]` I8\u002FI9: no implicit state, non-interactive auth\n- `[P1]` G6\u002FG9: precondition checks, fail-closed\n- `[P2]` G4\u002FG5\u002FG7: permission levels, PII redaction, batch limits\n\n#### Reserved Flags\n\n| Flag | Semantics | Notes |\n|------|-----------|-------|\n| `--agent` | JSON output (default) | Explicit override |\n| `--human` | Human-friendly output | Colors, tables, formatted |\n| `--brief` | One-paragraph identity | For sync into agent config |\n| `--help` | Full self-description JSON | Brief + commands + rules + skills + issue |\n| `--version` | Semver version string | |\n| `--yes` | Confirm destructive ops | Required for delete\u002Fdestroy |\n| `--dry-run` | Preview without executing | |\n| `--quiet` | Suppress stderr output | |\n| `--fields` | Filter output fields | Save tokens |\n\n### Level 3: Agent-Native (+ ecosystem -- 19 rules)\n\nGoal: CLI has identity, behavior contract, skill system, and feedback loop. Agent can learn the tool, extend its use, and report problems -- full closed-loop collaboration.\n\n**Agent Directory** -- tool identity and behavior contract\n- `[P1]` D12: `agent\u002Fbrief.md` exists\n- `[P1]` D13: `agent\u002Frules\u002F` has trigger.md, workflow.md, writeback.md\n- `[P1]` D17: agent\u002Frules\u002F*.md have YAML frontmatter (name, description)\n- `[P1]` D18: agent\u002Fskills\u002F*.md have YAML frontmatter (name, description)\n- `[P2]` D14: `agent\u002Fskills\u002F` directory + `skills` subcommand\n\n**Response Structure** -- inline context on every call\n- `[P1]` R1: Every response includes `rules[]` (full content from agent\u002Frules\u002F)\n- `[P1]` R2: Every response includes `skills[]` (name + description + command)\n- `[P1]` R3: Every response includes `issue` (feedback guide)\n\n**Meta** -- project-level integration\n- `[P2]` M1: AGENTS.md at project root\n- `[P2]` M2: Optional MCP tool schema export\n- `[P2]` M3: CHANGELOG.md marks breaking changes\n\n**Feedback** -- built-in issue system\n- `[P2]` F1: `issue` subcommand (create\u002Flist\u002Fshow)\n- `[P2]` F2: Structured submission with version\u002Fcontext\u002Fexit_code\n- `[P2]` F3: Categories: bug \u002F requirement \u002F suggestion \u002F bad-output\n- `[P2]` F4: Issues stored locally, no external service dependency\n- `[P2]` F5: `issue list` \u002F `issue show \u003Cid>` queryable\n- `[P2]` F6: Issues have status tracking (open\u002Fin-progress\u002Fresolved\u002Fclosed)\n- `[P2]` F7: Issue JSON has all required fields (id, type, status, message, created_at, updated_at)\n- `[P2]` F8: All issues have status field\n\n## Examples\n\n### Example 1: JSON Output (Agent Mode)\n\n```bash\n$ mycli list\n{\"result\": [{\"id\": 1, \"title\": \"Buy milk\", \"status\": \"todo\"}], \"rules\": [...], \"skills\": [...], \"issue\": \"...\"}\n```\n\n### Example 2: Structured Error\n\n```json\n{\n  \"error\": true,\n  \"code\": \"AUTH_EXPIRED\",\n  \"message\": \"Access token expired 2 hours ago\",\n  \"suggestion\": \"Run 'mycli auth refresh' to get a new token\"\n}\n```\n\n### Example 3: Exit Code Table\n\n```\n0   success         10  auth failed       20  resource not found\n1   general error   11  permission denied 30  conflict\u002Fprecondition\n2   param\u002Fusage error\n```\n\n## Quick Implementation Checklist\n\nImplement by layer -- each phase gets you the next certification level.\n\n**Phase 1: Agent-Friendly (core)**\n1. Default output is JSON -- no `--json` flag needed\n2. Error handler: `{ error, code, message, suggestion }` to stderr\n3. Exit codes: 0 success, 2 param error, 1 general\n4. stdout = data only, stderr = logs only\n5. Missing param -> structured error (never interactive)\n6. `--yes` guard on destructive operations\n7. Guardrails: reject secrets, path traversal, shell metacharacters\n\n**Phase 2: Agent-Ready (+ recommended)**\n8. `--help` returns structured JSON (help, commands[], rules[], skills[])\n9. `--brief` reads and outputs `agent\u002Fbrief.md` content\n10. `--human` flag switches to human-friendly format\n11. Reserved flags: --agent, --version, --dry-run, --quiet, --fields\n12. Exit codes: 20 not found, 30 conflict, 10 auth, 11 permission\n\n**Phase 3: Agent-Native (+ ecosystem)**\n13. Create `agent\u002F` directory: `brief.md`, `rules\u002Ftrigger.md`, `rules\u002Fworkflow.md`, `rules\u002Fwriteback.md`\n14. Every command response appends: rules[] + skills[] + issue\n15. `skills` subcommand: list all \u002F show one with full content\n16. `issue` subcommand for feedback (create\u002Flist\u002Fshow\u002Fclose\u002Ftransition)\n17. AGENTS.md at project root\n\n## Best Practices\n\n- Do: Default to JSON output so agents never need to add flags\n- Do: Include `suggestion` field in every error response\n- Do: Use the three-level certification model for incremental adoption\n- Do: Keep `agent\u002Fbrief.md` to one paragraph for token efficiency\n- Don't: Enter interactive mode on errors -- always exit immediately\n- Don't: Change JSON schema or error codes within the same version\n- Don't: Put logs or progress info on stdout -- use stderr only\n- Don't: Accept unknown flags silently -- reject with exit code 2\n\n## Common Pitfalls\n\n- **Problem:** CLI outputs human-readable text by default, breaking agent parsing\n  **Solution:** Make JSON the default output format; add `--human` flag for human-friendly mode\n\n- **Problem:** Errors reported in stdout with exit code 0\n  **Solution:** Always exit non-zero on failure and write structured error JSON to stderr\n\n- **Problem:** CLI prompts for missing input interactively\n  **Solution:** Return structured error with suggestion field and exit immediately\n\n## Related Skills\n\n- `@cli-best-practices` - General CLI design patterns (this skill focuses specifically on AI agent compatibility)\n\n## Additional Resources\n\n- [Agent CLI Spec Repository](https:\u002F\u002Fgithub.com\u002FChaosRealmsAI\u002Fagent-cli-spec)\n\n## Limitations\n- Use this skill only when the task clearly matches the scope described above.\n- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.\n- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.\n","# Agent-Friendly CLI Spec v0.1\n\nWhen building or modifying CLI tools, follow these rules to make them safe and\nreliable for AI agents to use.\n\n## Overview\n\nA comprehensive design specification for building AI-native CLI tools. It defines\n98 rules across three certification levels (Agent-Friendly, Agent-Ready, Agent-Native)\nwith prioritized requirements (P0\u002FP1\u002FP2). The spec covers structured JSON output,\nerror handling, input contracts, safety guardrails, exit codes, self-description,\nand a feedback loop via a built-in issue system.\n\n## When to Use This Skill\n\n- Use when building a new CLI tool that AI agents will invoke\n- Use when retrofitting an existing CLI to be agent-friendly\n- Use when designing command-line interfaces for automation pipelines\n- Use when auditing a CLI tool's compliance with agent-safety standards\n\n## Core Philosophy\n\n1. **Agent-first** -- default output is JSON; human-friendly is opt-in via `--human`\n2. **Agent is untrusted** -- validate all input at the same level as a public API\n3. **Fail-Closed** -- when validation logic itself errors, deny by default\n4. **Verifiable** -- every rule is written so it can be automatically checked\n\n## Layer Model\n\nThis spec uses two orthogonal axes:\n\n- **Layer** answers rollout scope: `core`, `recommended`, `ecosystem`\n- **Priority** answers severity: `P0`, `P1`, `P2`\n\nUse layers for migration and certification:\n\n- **core** -- execution contract: JSON, errors, exit codes, stdout\u002Fstderr, safety\n- **recommended** -- better machine UX: self-description, explicit modes, richer schemas\n- **ecosystem** -- agent-native integration: `agent\u002F`, `skills`, `issue`, inline context\n\nCertification maps to layers:\n\n- **Agent-Friendly** -- all `core` rules pass\n- **Agent-Ready** -- all `core` + `recommended` rules pass\n- **Agent-Native** -- all layers pass\n\n## How It Works\n\n### Step 1: Output Mode\n\nDefault is agent mode (JSON). Explicit flags to switch:\n\n```bash\n$ mycli list              # default = JSON output (agent mode)\n$ mycli list --human      # human-friendly: colored, tables, formatted\n$ mycli list --agent      # explicit agent mode (override config if needed)\n```\n\n- **Default (no flag)** -- JSON to stdout. Agent never needs to add a flag.\n- **--human** -- human-friendly format (colors, tables, progress bars)\n- **--agent** -- explicit JSON mode (useful when env\u002Fconfig overrides default)\n\n### Step 2: agent\u002F Directory Convention\n\nEvery CLI tool MUST have an `agent\u002F` directory at its project root. This is the\ntool's identity and behavior contract for AI agents.\n\n```\nagent\u002F\n  brief.md          # One paragraph: who am I, what can I do\n  rules\u002F            # Behavior constraints (auto-registered)\n    trigger.md      # When should an agent use this tool\n    workflow.md     # Step-by-step usage flow\n    writeback.md    # How to write feedback back\n  skills\u002F           # Extended capabilities (auto-registered)\n    getting-started.md\n```\n\n### Step 3: Four Levels of Self-Description\n\n1. **--brief** (business card, injected into agent config)\n2. **Every Command Response** (always-on context: data + rules + skills + issue)\n3. **--help** (full self-description: brief + commands + rules + skills + issue)\n4. **skills \\\u003Cname\\>** (on-demand deep dive into a specific skill)\n\n## Certification Requirements\n\nEach level includes all rules from the previous level.\nPriority tag `[P0]`=agent breaks without it, `[P1]`=agent works but poorly, `[P2]`=nice to have.\n\n### Level 1: Agent-Friendly (core -- 20 rules)\n\nGoal: CLI is a stable, callable API. Agent can invoke, parse, and handle errors.\n\n**Output** -- default is JSON, stable schema\n- `[P0]` O1: Default output is JSON. No `--json` flag needed\n- `[P0]` O2: JSON MUST pass `jq .` validation\n- `[P0]` O3: JSON schema MUST NOT change within same version\n\n**Error** -- structured, to stderr, never interactive\n- `[P0]` E1: Errors -> `{\"error\":true, \"code\":\"...\", \"message\":\"...\", \"suggestion\":\"...\"}` to stderr\n- `[P0]` E4: Error has machine-readable `code` (e.g. `MISSING_REQUIRED`)\n- `[P0]` E5: Error has human-readable `message`\n- `[P0]` E7: On error, NEVER enter interactive mode -- exit immediately\n- `[P0]` E8: Error codes are API contracts -- MUST NOT rename across versions\n\n**Exit Code** -- predictable failure signals\n- `[P0]` X3: Parameter\u002Fusage errors MUST exit 2\n- `[P0]` X9: Failures MUST exit non-zero -- never exit 0 then report error in stdout\n\n**Composability** -- clean pipe semantics\n- `[P0]` C1: stdout is for data ONLY\n- `[P0]` C2: logs, progress, warnings go to stderr ONLY\n\n**Input** -- fail fast on bad input\n- `[P1]` I4: Missing required param -> structured error, never interactive prompt\n- `[P1]` I5: Type mismatch -> exit 2 + structured error\n\n**Safety** -- protect against agent mistakes\n- `[P1]` S1: Destructive ops require `--yes` confirmation\n- `[P1]` S4: Reject `..\u002F..\u002F` path traversal, control chars\n\n**Guardrails** -- runtime input protection\n- `[P1]` G1: Unknown flags rejected with exit 2\n- `[P1]` G2: Detect API key \u002F token patterns in args, reject execution\n- `[P1]` G3: Reject sensitive file paths (*.env, *.key, *.pem)\n- `[P1]` G8: Reject shell metacharacters in arguments (; | && $())\n\n### Level 2: Agent-Ready (+ recommended -- 59 rules)\n\nGoal: CLI is self-describing, well-named, and pipe-friendly. Agent discovers capabilities and chains commands without trial and error.\n\n**Self-Description** -- agent discovers what CLI can do\n- `[P1]` D1: `--help` outputs structured JSON with `commands[]`\n- `[P1]` D3: Schema has required fields (help, commands)\n- `[P1]` D4: All parameters have type declarations\n- `[P1]` D7: Parameters annotated as required\u002Foptional\n- `[P1]` D9: Every command has a description\n- `[P1]` D11: `--help` outputs JSON with help, rules, skills, commands\n- `[P1]` D15: `--brief` outputs `agent\u002Fbrief.md` content\n- `[P1]` D16: Default JSON (agent mode), `--human` for human-friendly\n- `[P2]` D2\u002FD5\u002FD6\u002FD8\u002FD10: per-command help, enums, defaults, output schema, version\n\n**Input** -- unambiguous calling convention\n- `[P1]` I1: All flags use `--long-name` format\n- `[P1]` I2: No positional argument ambiguity\n- `[P2]` I3\u002FI6\u002FI7: --json-input, boolean --no-X, array params\n\n**Error**\n- `[P1]` E6: Error includes `suggestion` field\n- `[P2]` E2\u002FE3: errors to stderr, error JSON valid\n\n**Safety**\n- `[P1]` S8: `--sanitize` flag for external input\n- `[P2]` S2\u002FS3\u002FS5\u002FS6\u002FS7: default deny, --dry-run, no auto-update, destructive marking\n\n**Exit Code**\n- `[P1]` X1: 0 = success\n- `[P2]` X2\u002FX4-X8: 1=general, 10=auth, 11=permission, 20=not-found, 30=conflict\n\n**Composability**\n- `[P1]` C6: No interactive prompts in pipe mode\n- `[P2]` C3\u002FC4\u002FC5\u002FC7: pipe-friendly, --quiet, pipe chain, idempotency\n\n**Naming** -- predictable flag conventions\n- `[P1]` N4: Reserved flags (--agent, --human, --brief, --help, --version, --yes, --dry-run, --quiet, --fields)\n- `[P2]` N1\u002FN2\u002FN3\u002FN5\u002FN6: consistent naming, kebab-case, max 3 levels, --version semver\n\n**Guardrails**\n- `[P1]` I8\u002FI9: no implicit state, non-interactive auth\n- `[P1]` G6\u002FG9: precondition checks, fail-closed\n- `[P2]` G4\u002FG5\u002FG7: permission levels, PII redaction, batch limits\n\n#### Reserved Flags\n\n| Flag | Semantics | Notes |\n|------|-----------|-------|\n| `--agent` | JSON output (default) | Explicit override |\n| `--human` | Human-friendly output | Colors, tables, formatted |\n| `--brief` | One-paragraph identity | For sync into agent config |\n| `--help` | Full self-description JSON | Brief + commands + rules + skills + issue |\n| `--version` | Semver version string | |\n| `--yes` | Confirm destructive ops | Required for delete\u002Fdestroy |\n| `--dry-run` | Preview without executing | |\n| `--quiet` | Suppress stderr output | |\n| `--fields` | Filter output fields | Save tokens |\n\n### Level 3: Agent-Native (+ ecosystem -- 19 rules)\n\nGoal: CLI has identity, behavior contract, skill system, and feedback loop. Agent can learn the tool, extend its use, and report problems -- full closed-loop collaboration.\n\n**Agent Directory** -- tool identity and behavior contract\n- `[P1]` D12: `agent\u002Fbrief.md` exists\n- `[P1]` D13: `agent\u002Frules\u002F` has trigger.md, workflow.md, writeback.md\n- `[P1]` D17: agent\u002Frules\u002F*.md have YAML frontmatter (name, description)\n- `[P1]` D18: agent\u002Fskills\u002F*.md have YAML frontmatter (name, description)\n- `[P2]` D14: `agent\u002Fskills\u002F` directory + `skills` subcommand\n\n**Response Structure** -- inline context on every call\n- `[P1]` R1: Every response includes `rules[]` (full content from agent\u002Frules\u002F)\n- `[P1]` R2: Every response includes `skills[]` (name + description + command)\n- `[P1]` R3: Every response includes `issue` (feedback guide)\n\n**Meta** -- project-level integration\n- `[P2]` M1: AGENTS.md at project root\n- `[P2]` M2: Optional MCP tool schema export\n- `[P2]` M3: CHANGELOG.md marks breaking changes\n\n**Feedback** -- built-in issue system\n- `[P2]` F1: `issue` subcommand (create\u002Flist\u002Fshow)\n- `[P2]` F2: Structured submission with version\u002Fcontext\u002Fexit_code\n- `[P2]` F3: Categories: bug \u002F requirement \u002F suggestion \u002F bad-output\n- `[P2]` F4: Issues stored locally, no external service dependency\n- `[P2]` F5: `issue list` \u002F `issue show \u003Cid>` queryable\n- `[P2]` F6: Issues have status tracking (open\u002Fin-progress\u002Fresolved\u002Fclosed)\n- `[P2]` F7: Issue JSON has all required fields (id, type, status, message, created_at, updated_at)\n- `[P2]` F8: All issues have status field\n\n## Examples\n\n### Example 1: JSON Output (Agent Mode)\n\n```bash\n$ mycli list\n{\"result\": [{\"id\": 1, \"title\": \"Buy milk\", \"status\": \"todo\"}], \"rules\": [...], \"skills\": [...], \"issue\": \"...\"}\n```\n\n### Example 2: Structured Error\n\n```json\n{\n  \"error\": true,\n  \"code\": \"AUTH_EXPIRED\",\n  \"message\": \"Access token expired 2 hours ago\",\n  \"suggestion\": \"Run 'mycli auth refresh' to get a new token\"\n}\n```\n\n### Example 3: Exit Code Table\n\n```\n0   success         10  auth failed       20  resource not found\n1   general error   11  permission denied 30  conflict\u002Fprecondition\n2   param\u002Fusage error\n```\n\n## Quick Implementation Checklist\n\nImplement by layer -- each phase gets you the next certification level.\n\n**Phase 1: Agent-Friendly (core)**\n1. Default output is JSON -- no `--json` flag needed\n2. Error handler: `{ error, code, message, suggestion }` to stderr\n3. Exit codes: 0 success, 2 param error, 1 general\n4. stdout = data only, stderr = logs only\n5. Missing param -> structured error (never interactive)\n6. `--yes` guard on destructive operations\n7. Guardrails: reject secrets, path traversal, shell metacharacters\n\n**Phase 2: Agent-Ready (+ recommended)**\n8. `--help` returns structured JSON (help, commands[], rules[], skills[])\n9. `--brief` reads and outputs `agent\u002Fbrief.md` content\n10. `--human` flag switches to human-friendly format\n11. Reserved flags: --agent, --version, --dry-run, --quiet, --fields\n12. Exit codes: 20 not found, 30 conflict, 10 auth, 11 permission\n\n**Phase 3: Agent-Native (+ ecosystem)**\n13. Create `agent\u002F` directory: `brief.md`, `rules\u002Ftrigger.md`, `rules\u002Fworkflow.md`, `rules\u002Fwriteback.md`\n14. Every command response appends: rules[] + skills[] + issue\n15. `skills` subcommand: list all \u002F show one with full content\n16. `issue` subcommand for feedback (create\u002Flist\u002Fshow\u002Fclose\u002Ftransition)\n17. AGENTS.md at project root\n\n## Best Practices\n\n- Do: Default to JSON output so agents never need to add flags\n- Do: Include `suggestion` field in every error response\n- Do: Use the three-level certification model for incremental adoption\n- Do: Keep `agent\u002Fbrief.md` to one paragraph for token efficiency\n- Don't: Enter interactive mode on errors -- always exit immediately\n- Don't: Change JSON schema or error codes within the same version\n- Don't: Put logs or progress info on stdout -- use stderr only\n- Don't: Accept unknown flags silently -- reject with exit code 2\n\n## Common Pitfalls\n\n- **Problem:** CLI outputs human-readable text by default, breaking agent parsing\n  **Solution:** Make JSON the default output format; add `--human` flag for human-friendly mode\n\n- **Problem:** Errors reported in stdout with exit code 0\n  **Solution:** Always exit non-zero on failure and write structured error JSON to stderr\n\n- **Problem:** CLI prompts for missing input interactively\n  **Solution:** Return structured error with suggestion field and exit immediately\n\n## Related Skills\n\n- `@cli-best-practices` - General CLI design patterns (this skill focuses specifically on AI agent compatibility)\n\n## Additional Resources\n\n- [Agent CLI Spec Repository](https:\u002F\u002Fgithub.com\u002FChaosRealmsAI\u002Fagent-cli-spec)\n\n## Limitations\n- Use this skill only when the task clearly matches the scope described above.\n- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.\n- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.\n",{"name":5,"risk":24,"source":25,"date_added":26,"description":7},"safe","https:\u002F\u002Fgithub.com\u002FChaosRealmsAI\u002Fagent-cli-spec","2026-03-15","main","87afe6ce96220accf50a345dea46784c8c9f210e","YYC3-Skills-技能包\u002FYYC3-Skills-技能包-antigravity-awesome-skills\u002Fplugins\u002Fantigravity-awesome-skills-claude\u002Fskills\u002Fai-native-cli",0,"https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base","https:\u002F\u002Fkb.yyc3.top\u002F","https:\u002F\u002Fgithub.com\u002FYYC-Cube","万象归元于云枢 | 深栈智启新纪元","2026-11-10 10:37:32",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill ai-native-cli",{"tool":42,"command":43},"bunx","bunx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill ai-native-cli",{"tool":45,"command":46},"pnpm","pnpm dlx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill ai-native-cli","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-Skills-技能包\u002FYYC3-Skills-技能包-antigravity-awesome-skills\u002Fplugins\u002Fantigravity-awesome-skills-claude\u002Fskills\u002Fai-native-cli\u002FSKILL.md",true,[],[51,58,65,72,79,86],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":19},"01M7CTDWHERCXPQCP0K03A86AS","vercel-sandbox","Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include \"Vercel Sandbox browser\", \"microVM Chrome\", \"agent-browser in sandbox\", \"browser automation on Vercel\", or any task requiring Chrome in a Vercel Sandbox.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fvercel-sandbox\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":19},"01M7CTD8D2BNJY54S65M47HEJA","slack","Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include \"check my Slack\", \"what channels have unreads\", \"send a message to\", \"search Slack for\", \"extract from Slack\", \"find who said\", or any task requiring programmatic Slack interaction.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fslack\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7CTCSSB8ZP3Q58C3YZDZK1W","electron","Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include \"automate Slack app\", \"control VS Code\", \"interact with Discord app\", \"test this Electron app\", \"connect to desktop app\", or any task requiring automation of a native Electron application.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Felectron\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M7CTCBJY76688V4DCZGWFHGQ","dogfood","Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to \"dogfood\", \"QA\", \"exploratory test\", \"find issues\", \"bug hunt\", \"test this app\u002Fsite\u002Fplatform\", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fdogfood\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M7CTC4H0FFGVEAQR9PA4GA1A","agent-browser","Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to \"open a website\", \"fill out a form\", \"click a button\", \"take a screenshot\", \"scrape data from a page\", \"test this web app\", \"login to a site\", \"automate browser actions\", or any task requiring programmatic web interaction.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fagent-browser\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M7CTBKVFYBG78FDSABSDJTTQ","shadcn","Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI, including chat interfaces. Provides project context, component docs, and usage examples. Applies when working with shadcn\u002Fui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for \"shadcn init\", \"create an app with --preset\", \"switch to --preset\", and for building a design system or applying a DESIGN.md or brand spec to shadcn\u002Fui.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-UI-界面组件\u002FYYC3-UI-界面组件-ui\u002Fskills\u002Fshadcn\u002FSKILL.md",[],[]]