The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted data from the repository to inform agent decisions.
Ingestion points: The skill reads data from Git diffs, commit messages, issue tracker documentation, and specification files located in docs/, specs/, or .scratch/.
Boundary markers: There are no explicit markers or delimiters defined in the instructions to isolate the interpolated external content from the agent's core instructions.
Capability inventory: The skill is capable of executing shell commands via git and spawning sub-agents using the Agent tool.
Sanitization: No sanitization or validation of the ingested file content or diff output is performed prior to inclusion in sub-agent prompts.
[COMMAND_EXECUTION]: The skill uses shell commands (git diff and git log) to analyze code changes. These operations are restricted to read-only repository metadata and are consistent with the skill's stated purpose of performing code reviews.

review