The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes source code files to generate and validate documentation, creating a surface for indirect prompt injection if the code contains malicious instructions within comments or docstrings.
Ingestion points: The skill reads source code files during the 'Analyze' and 'Document' phases as defined in the SKILL.md workflow.
Boundary markers: There are no explicit instructions to the agent to disregard or escape natural language instructions found within the code being documented.
Capability inventory: The skill possesses the capability to write to the filesystem (generating documentation) and execute local commands through the validation step.
Sanitization: No sanitization of code examples is performed before they are passed to validation tools.
[DYNAMIC_EXECUTION]: The workflow incorporates automated testing of documentation examples, which involves executing code snippets at runtime.
Evidence: SKILL.md specifies the use of python -m doctest and pytest --doctest-modules to verify that examples in docstrings are functional and produce expected results.
[COMMAND_EXECUTION]: The skill utilizes several standard development and documentation tools via the command line for linting and validation.
Evidence: SKILL.md and references/coverage-reports.md include instructions to run tsc --noEmit, npx @redocly/cli lint, pydocstyle, and interrogate.
[EXTERNAL_DOWNLOADS]: The skill references documentation and configuration files from external sites and well-known service providers.
Evidence: SKILL.md provides a link to https://jeffallan.github.io/claude-skills/skills/quality/code-documenter/ for further documentation.
Evidence: references/documentation-systems.md and references/interactive-api-docs.md include script tags and links for external libraries such as Algolia (cdn.jsdelivr.net) and Redocly (cdn.redoc.ly) to support documentation search and portal functionality.

code-documenter