The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of code diffs and file contents which are then processed by a sub-agent.
Ingestion points: SKILL.md (Step 1) gathers data using git diff or reading file contents.
Boundary markers: references/reviewer-prompt.md uses structural delimiters like markdown headers but lacks strict escaping or instructions to ignore embedded adversarial prompts within the code.
Capability inventory: The skill uses git diff and spawns a sub-agent with tool-access via create_agent.
Sanitization: No sanitization is performed on the code contents before they are passed into the review prompt.
[COMMAND_EXECUTION]: The skill invokes system commands to retrieve repository state.
Evidence: SKILL.md (Step 1) specifies the execution of git diff main...HEAD to identify changes for review. This is an expected functional component of a code review tool.

interrogate