{
  "markdown": "# Convex Agent Skills\n\nAgent skills for common Convex workflows.\n\n## License\n\nUnless otherwise noted, all original content in this repository is licensed\nunder the [Apache License 2.0](LICENSE). This includes the skill instructions,\nreference documentation, and source code in revisions that include this license.\nIt does not retroactively change the licensing of earlier revisions.\n\nThird-party materials remain under their respective licenses. See\n[THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md) for details.\n\n## Install\n\n```bash\n# Choose which skills you want\nnpx skills add get-convex/agent-skills\n\n# OR Install all skills\nnpx skills add get-convex/agent-skills --all\n```\n\n## Usage\n\nSkills are applied automatically when the agent determines they're relevant. How\nyou manually invoke them depends on your tool:\n\n| Tool                     | Manual invocation |\n| ------------------------ | ----------------- |\n| Cursor                   | `/skill-name`     |\n| VS Code (GitHub Copilot) | `/skill-name`     |\n| Claude Code              | `/skill-name`     |\n| Windsurf                 | `@skill-name`     |\n| Codex (OpenAI)           | `$skill-name`     |\n\nFor example, to kick off auth setup in Cursor or Claude Code:\n\n```\n/convex-setup-auth\n```\n\nIn Windsurf:\n\n```\n@convex-setup-auth\n```\n\n## Available Skills\n\n- `convex` - Top-level entry point for Convex work. Recommends installing Convex\n  AI files, then routes to a more specific Convex skill.\n- `convex-quickstart` - Start a new Convex project or add Convex to an existing\n  app.\n- `convex-setup-auth` - Set up authentication for a Convex app.\n- `convex-create-component` - Build a reusable Convex component with clear\n  boundaries.\n- `convex-migration-helper` - Plan and run Convex migrations safely.\n- `convex-performance-audit` - Investigate Convex performance problems and\n  bottlenecks.\n\n## Skill Philosophy\n\nSkills in this repo should be laser-focused on a specific task or workflow.\n\nA good skill helps an agent take action, for example:\n\n- set up authentication\n- create a component\n- perform a migration\n- diagnose performance issues\n\nA skill should not exist just to provide generic background information. If\ncontent is mostly reference material, it should usually live in documentation,\nnot as a standalone skill.\n\nReference material is still useful inside a skill, but only when it helps the\nagent complete a concrete task.\n\n## Contributing\n\nBefore contributing, review the core Agent Skills docs:\n\n- [Overview](https://agentskills.io/home)\n- [What are skills?](https://agentskills.io/what-are-skills)\n- [Specification](https://agentskills.io/specification)\n- [Optimizing skill descriptions](https://agentskills.io/skill-creation/optimizing-descriptions)\n- [Evaluating skill output quality](https://agentskills.io/skill-creation/evaluating-skills)\n\nIf your skill bundles scripts, also read\n[Using scripts in skills](https://agentskills.io/skill-creation/using-scripts).\n\nInstall dependencies once with `npm install`, then use:\n\n- `npm run format` to rewrite files\n- `npm run format-check` to verify formatting before opening a PR\n\n### Validating Skills\n\nValidate skills by trying to use them in a realistic temp project, not just by\nreading them.\n\n- Push the skill as far as possible with an agent in a throwaway directory\n- If a human must intervene, ask explicitly for the exact action needed and then\n  continue\n- Record what worked, where the agent got stuck, and what confused the flow\n- Feed those learnings back into the skill so the next run is better\n- For UI-facing skills such as auth setup, validate the actual browser flow\n  during skill development, not just code generation or a successful build\n\n### Testing with Anthropic's Skill Creator\n\nFor a more rigorous approach, use Anthropic's\n[skill-creator](https://github.com/anthropics/skills/tree/main/skills/skill-creator)\nskill. It provides a structured loop for testing and iterating on skills:\n\n1. Draft or edit a skill\n2. Define realistic test prompts and run the skill against them (with and\n   without the skill as a baseline)\n3. Grade the outputs with assertions and review them in a browser-based viewer\n4. Improve the skill based on feedback, then repeat\n\nIt also includes a description optimizer that tunes the skill's frontmatter\ndescription for better triggering accuracy across different phrasings.\n\nInstall it in Claude Code:\n\n```bash\n/install-skill https://github.com/anthropics/skills/tree/main/skills/skill-creator\n```\n\nThis is especially useful when a skill is complex enough that reading it alone\nwon't tell you if it actually works well in practice.\n\nEach skill follows the\n[Agent Skills open standard](https://github.com/anthropics/skills):\n\n1. Create a directory under `skills/` with the skill name\n2. Add a `SKILL.md` file with YAML frontmatter:\n   ```yaml\n   ---\n   name: skill-name\n   description: Brief description\n   ---\n   ```\n3. Include comprehensive examples with bad/good patterns\n4. Add a checklist at the end of each skill\n5. Update this root `README.md` whenever skills are added, removed, renamed, or\n   substantially repositioned\n",
  "bytes": 5103,
  "sha": "793eee945acf06a4ac06e8c88581175a0211dd89bb1f637852c29f80eb93ac84",
  "repo_slug": "get-convex/agent-skills",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_get_convex_agent_skills_convex_reviewer_a0fc4465/readme"
}