The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses rg (ripgrep) to discover existing configurations and npm / npx to install and run the metalint orchestrator and delegated linters.
[EXTERNAL_DOWNLOADS]: Fetches the metalint package and associated linting tools (ESLint, Stylelint, HTMLHint) from the official npm registry.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from repository files to determine its configuration, which represents a standard attack surface for indirect instructions.
Ingestion points: Reads package.json, AGENTS.md, and various linter configuration files (e.g., eslint.config.js).
Boundary markers: No specific delimiters or warnings for embedded instructions are defined in the instructions.
Capability inventory: Includes file system write access (for creating/updating configuration files and package scripts) and command execution capabilities (npm, npx, rg).
Sanitization: The workflow does not explicitly define sanitization steps for the ingested repository content.

metalint