The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to process user-provided source code and documentation, which creates a surface for indirect prompt injection. Malicious instructions hidden within the code being reviewed could attempt to influence the agent's logic or output.
Ingestion points: User-supplied source code, docstrings, and comments provided for review across all programming languages.
Boundary markers: The skill instructions do not specify any delimiters or safety markers to isolate user-provided code from the agent's operational instructions.
Capability inventory: The skill utilizes text generation and internal search tools (ripgrep) to query its reference documentation. It does not request access to high-risk capabilities such as arbitrary shell execution, network connections, or unauthorized file system access.
Sanitization: No sanitization, escaping, or validation mechanisms for user-provided input are described in the skill configuration.

code-commenting