The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes local repository documentation and source code to verify facts, paths, and symbols.
Ingestion points: Markdown files and TypeScript/JavaScript source files analyzed by scripts/docs-probe.mts, scripts/arch-lint.mts, and scripts/check-architecture-page.mts.
Boundary markers: The skill instructs the agent to perform final reviews using standing criteria and existing review skills.
Capability inventory: File reading across the repository and localized file writing for rendering architecture maps via scripts/render-architecture-map.mts.
Sanitization: Employs structural parsing (AST analysis via oxc-parser and markdown tokenization) rather than execution of analyzed content.
[EXTERNAL_DOWNLOADS]: The architecture linting scripts utilize the oxc-parser package, which is a well-known, high-performance library for parsing source code. The documentation provides standard instructions for installing this dependency.
[COMMAND_EXECUTION]: The skill includes multiple TypeScript scripts designed to be executed in a project's local development environment using the Bun runtime to automate documentation quality checks and architecture verification.

docs-discipline