The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git diffs and local repository documentation (such as AGENTS.md and project docs) which could contain hidden instructions intended to influence the agent's refactoring logic or validation steps. * Ingestion points: Output from git diff commands and content from repository-level instruction files (SKILL.md). * Boundary markers: The skill instructs the agent to read local documentation to contextualize changes, but it does not specify clear delimiters or ignore-instructions for the code under review. * Capability inventory: Includes reading files (git diff), writing behavior-preserving fixes, and executing local validation tools (tests, linters). * Sanitization: Implements a review process that aggregates and filters findings before applying fixes, limiting changes to behavior-preserving actions.
[COMMAND_EXECUTION]: The skill executes shell commands for git operations (git diff, git diff --cached) and validation tasks (tests, typecheck, linting). These commands are executed within the local environment and are targeted at the files being reviewed.

simplify-code