The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external codebase content (source code, configuration files, and comments) to generate documentation, creating a surface for indirect prompt injection.
Ingestion points: Project source files are read using Python's ast module and glob patterns as demonstrated in resources/implementation-playbook.md.
Boundary markers: The skill does not define specific delimiters or instructions to ignore potential malicious prompts embedded within code comments or docstrings.
Capability inventory: The skill includes capabilities for reading local files, generating new documentation files, and provides examples for CI/CD pipeline integration.
Sanitization: While the skill includes a safety instruction to avoid exposing secrets, it lacks specific sanitization logic to handle adversarial instructions hidden in processed code data.
[EXTERNAL_DOWNLOADS]: The documentation CI/CD example in resources/implementation-playbook.md references standard tools and services.
The GitHub Actions workflow utilizes well-known actions such as actions/checkout, actions/setup-python, and peaceiris/actions-gh-pages.
The playbook mentions the installation of @redocly/cli via NPM, which is a widely recognized tool for OpenAPI documentation.
[SAFE]: The skill demonstrates security best practices by including a dedicated 'Safety' section that instructs the agent to avoid exposing secrets, internal URLs, or sensitive data during the documentation generation process.

documentation-generation-doc-generate