The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands specifically restricted to the prek prefix to install and verify hook configurations (e.g., prek install, prek run --all-files).
[EXTERNAL_DOWNLOADS]: The skill recommends hooks from well-known community sources, such as the ruff-pre-commit repository by Astral (a known software vendor in the Python ecosystem).
[PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface by reading untrusted project configuration files to determine the appropriate hooks to suggest.
Ingestion points: Reads project-specific files including Cargo.toml, pyproject.toml, package.json, go.mod, Gemfile, Dockerfile, and existing configuration files (prek.toml, .pre-commit-config.yaml).
Boundary markers: None identified in the instructions for separating ingested file content from system instructions.
Capability inventory: Access to Bash (restricted to prek), Read, Write, Edit, Glob, and Grep tools.
Sanitization: No explicit sanitization or validation of the content of the read files before processing.

prek