The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard git commands (git diff, git log, git rev-parse) to analyze the repository state and changes. These commands are restricted to local repository information and are necessary for its primary function.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, specifically code diffs and content from an issue tracker, which could contain instructions meant to mislead the AI.
Ingestion points: The skill ingests git diff output and fetches issue data via workflows described in docs/agents/issue-tracker.md to perform its review.
Boundary markers: While the skill uses separate prompts for sub-agents to isolate the analysis axes, it does not specify explicit delimiters or "ignore embedded instructions" warnings for the ingested diff/issue content within those sub-agent prompts.
Capability inventory: The skill has the capability to spawn sub-agents via the Agent tool and read repository files.
Sanitization: The instructions do not mention sanitization, escaping, or validation of the external content before it is interpolated into the prompts for the sub-agents.

code-review