The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard git commands to identify commit ranges for review. These commands are restricted to repository metadata and do not access sensitive system files.
Evidence: SKILL.md uses git rev-parse HEAD and git log --oneline to manage the code review process.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and review project files and plans, which creates a surface for indirect prompt injection if those files contain malicious instructions.
Ingestion points: Reads git commit history and project documentation files like docs/plans/deployment-plan.md (specified in SKILL.md).
Boundary markers: None present; the skill does not explicitly instruct the subagent to ignore instructions embedded within the code being reviewed.
Capability inventory: Executes local git commands and dispatches a subagent (superpowers:code-reviewer).
Sanitization: None; the content of the files is passed directly to the review process. This is an inherent risk for code-analysis tools and is handled at the platform level.

requesting-code-review