The Agent Skills Directory

[SAFE]: A comprehensive security audit of the core skill instructions and all language-specific references (Python, JavaScript/TypeScript, and GitHub Workflows) found no evidence of malicious patterns, credential theft, or unauthorized persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to analyze untrusted code changes. 1. Ingestion points: Source code and configuration files are ingested via the Read, Grep, and Glob tools (SKILL.md). 2. Boundary markers: No explicit delimiters are specified to isolate the code from the instructions. 3. Capability inventory: The skill is restricted to read-only tools and text-based reporting; it has no access to network, file-writing, or code execution tools (SKILL.md). 4. Sanitization: No sanitization of ingested code content is performed. Given the lack of dangerous capabilities, the risk of ingested data influencing the system beyond the specific correctness report is minimal.

code-review