The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data from git diff outputs, creating a surface for indirect prompt injection where malicious code changes could attempt to influence the agent's review behavior.
Ingestion points: Diff outputs from git diff and contents of source files in the repository.
Boundary markers: Not explicitly defined in the skill instructions; the workflow refers to an external template for the report structure.
Capability inventory: Reading repository data via git, executing the local script diff_stats.py, and writing reports to the .reviews/ directory.
Sanitization: No explicit sanitization or filtering of the diff content is specified in the skill workflow.
[COMMAND_EXECUTION]: The skill executes local commands including git diff and a Python script diff_stats.py. These operations are consistent with the skill's stated purpose of reviewing code changes and do not involve unauthorized actions or privilege escalation.

code-review-md