メニュー
Root index of x-cmd skill0 sub-skills. Defines the OKR-style agent workflow (goal → rule-verified results → execute), skill discovery, and agent tooling preferences. Style: principle-first, concise, delegate specifics to authoritative external sources.
Browser automation via Chrome/Chromium CDP — open, snapshot, click, screenshot. For testing web apps, mobile layouts, and automated interactions without Playwright/Puppeteer.
Lovable MCP tool usage — deploy projects, message agent, test with agent-browser. Minimize credits by delegating to local repo.
Typed knowledge graph via TSV append-only log — entity CRUD, directed relations, schema validation, multi-hop reasoning. CLI via `x ondb`, protocol readable by AWK/Python/JS/SQLite.
Prompt engineering conventions for x-cmd — reuse via template variables, structure rules, safety enforcement patterns.
Writing conventions for skill0 documents — pyramid structure, line limits, and layout rules.
Multi-phase project assessment with scoring. Triggered by "/assess" to evaluate project health across dimensions and generate a comprehensive report.
| name | skill0 |
| description | Root index of x-cmd skill0 sub-skills. Defines the OKR-style agent workflow (goal → rule-verified results → execute), skill discovery, and agent tooling preferences. Style: principle-first, concise, delegate specifics to authoritative external sources. |
Skill0 is both the spec and the live prompts of x-cmd's AI system. It defines how x-cmd agents operate and is loaded as real prompts at runtime — not documentation, but the actual instructions agents execute.
Content philosophy: focus on stable principles. Avoid volatile info (API details, version-specific behavior, tutorials). Curate vetted authoritative links that agents can fetch and use at runtime. See skill0-writer for conventions including link curation rules.
x rule and x testSee the rule sub-skill, or x rule -h for the shell-integrated workflow.
x skill — x-cmd's curated, human-vetted skill catalogx clawhub — global skill registry. Caution: free upload, MUST check x clawhub skill moderate <name> for auto-generated safety report -- also noted it is AI-generated, not human-reviewed.Execution preferences
x roadmap (project management), x cron (scheduled tasks), x agent job (background agents), x ondb (ontology), x wiki / x llmwiki (wiki). Run x [mod] --help or check the corresponding sub-skill.
Unlike vendor-specific mechanisms, these are public, CLI-manageable, and cross-vendor.