The Agent Skills Directory

[PROMPT_INJECTION]: The skill incorporates safety instructions to mitigate indirect prompt injection. 1. Ingestion points: The agent reads file contents, comments, and git history as instructed in SKILL.md. 2. Boundary markers: The skill instructs the agent to clearly delimit code snippets as data. 3. Capability inventory: The skill is intended to edit files and execute tests. 4. Sanitization: It explicitly directs the agent to ignore any instructions embedded in the code or git data.
[COMMAND_EXECUTION]: The skill uses git diff to identify modified files for its tasks. This is a standard and legitimate use of version control tools for task scoping.
[DATA_EXFILTRATION]: No evidence of data exfiltration was found. The skill operates within the local repository context using file paths and metadata necessary for the requested simplification.

code-simplifier