The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/review_diff.py uses subprocess.run to execute local git commands, including git diff and git rev-parse. This allows the tool to automate the extraction of changes for review.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted data from git diffs which may contain malicious instructions embedded by external contributors. 1. Ingestion points: scripts/review_diff.py reads data directly from the output of git diff. 2. Boundary markers: The script interpolates raw diff content into a markdown report without using clear delimiters or security warnings for the AI agent. 3. Capability inventory: The skill has the capability to execute git via subprocess and write files to the local file system (defaulting to /tmp/review.md). 4. Sanitization: There is no validation or sanitization of the content retrieved from the diff before it is processed by the agent.

code-reviewer