The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided plans and design documents as well as contents of the local codebase. While this constitutes an attack surface for indirect prompt injection, the skill's functionality is limited to asking questions and providing recommendations to the user, with no autonomous execution of commands based on ingested data. No boundary markers or sanitization are explicitly defined in the prompt, but the limited capability tier results in a safe assessment.
[COMMAND_EXECUTION]: The skill instructs the agent to 'explore the codebase' to answer questions. This is a standard operation for developer-oriented AI agents and does not involve arbitrary or malicious command execution.
[DATA_EXFILTRATION]: There are no network operations, hardcoded credentials, or patterns suggesting the exfiltration of sensitive information from the codebase or user input.

grill-me