The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard shell commands to manage code review metadata and examine changes.
Evidence: Uses git rev-parse to identify commit SHAs and git log with pipe sequences (grep, head, awk) to locate specific task markers in history.
Evidence: The review subagent is instructed to use git diff to analyze code changes between the identified base and head commits.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface where instructions embedded in project documentation or code changes could influence the reviewer agent's behavior.
Ingestion points: The subagent template in code-reviewer.md ingests external data through {PLAN_OR_REQUIREMENTS}, {DESCRIPTION}, and {PLAN_REFERENCE}.
Boundary markers: The template does not utilize specific delimiters or instructions to ignore potential commands embedded in the review material.
Capability inventory: The subagent has the capability to execute filesystem read commands via git diff.
Sanitization: No sanitization or filtering of the input plan text or code diffs is performed prior to prompt interpolation.

requesting-code-review