The Agent Skills Directory

[SAFE]: The skill uses standard Git commands like git diff and git log to analyze branch differences. These operations are read-only and occur within the local repository context, posing no risk of data exfiltration or unauthorized system changes.
[SAFE]: The instructions include mandatory safety constraints that explicitly prohibit the agent from applying auto-fixes, making commits, or creating pull requests, ensuring the agent functions strictly as a text-only reviewer.
[SAFE]: Indirect Prompt Injection Risk Assessment: 1. Ingestion points: Git diff content processed by the agent (SKILL.md). 2. Boundary markers: No explicit delimiters are used for the code content. 3. Capability inventory: Limited to read-only Git commands (SKILL.md). 4. Sanitization: None provided. The risk is negligible as the agent only generates text feedback and lacks the capability to execute the code or perform network operations with the processed data.

code-review