{
  "markdown": "# Django AI Skills\n\nPortable Django expertise for Claude Code, Codex, Cursor, OpenCode, and tools\nthat implement the [Agent Skills specification](https://agentskills.io/).\n\nThe repository has one canonical authoring tree under `skills/`. Host manifests\nand package-local projections are generated from `plugins/catalog.json`, so the\nsame behavior is delivered without maintaining separate copies by hand.\n\n## Included capabilities\n\n| ID | Capability |\n| --- | --- |\n| `django-expert` | Django backend, ORM, DRF, security, testing, and performance guidance |\n| `django-celery-expert` | Reliable Celery task design, retries, scheduling, and operations |\n| `cdrf-expert` | DRF class selection, lifecycle tracing, MRO, and safe override guidance |\n| `django-safe-migration` | PostgreSQL-aware, zero-downtime Django migration guidance |\n| `django-reviewer` | Report-first Django and Python review for recent or user-selected changes |\n\nSee [installation and lifecycle instructions](docs/installation.md) for each\nhost and the [compatibility matrix](docs/compatibility.md) for the validation\nlevel of every adapter.\n\n## Repository layout\n\n```text\nskills/                         canonical Agent Skills\nplugins/catalog.json            canonical metadata and host declarations\nplugins/<id>/                   self-contained generated plugin packages\n.claude-plugin/                 Claude marketplace\n.agents/plugins/                Codex marketplace\n.cursor-plugin/                 Cursor marketplace\n.opencode/plugins/              OpenCode skills-path adapter\nscripts/                        generation, validation, and isolated smoke checks\n```\n\nPackage manifests and package-local skills are generated outputs. Contributors\nshould read [AGENTS.md](AGENTS.md) before changing the catalog or skill content.\n\n## Local validation\n\n```bash\npython scripts/generate_adapters.py --check\npython scripts/validate_plugins.py\npython -m unittest discover -s tests -p 'test_*.py'\n```\n\nThe smoke runner accepts `claude`, `codex`, `cursor`, `opencode`, and\n`agent-skills` targets and always installs into a temporary location:\n\n```bash\npython scripts/smoke_plugins.py --target cursor\n```\n\nThese skills can complement the\n[django-ai-boost MCP server](https://github.com/vintasoftware/django-ai-boost);\nneither project requires the other.\n\n## License\n\nMIT © Vinta Software\n",
  "bytes": 2352,
  "sha": "34f0d9cada873a402e39b83c6073c1df284bcc6f7ed94abd8f5fb56c7f00dd61",
  "repo_slug": "vintasoftware/django-ai-plugins",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_vintasoftware_django_ai_plugins_django_e_5653cf10/readme"
}