The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources, creating an attack surface for indirect prompt injection.\n
Ingestion points: The skill reads data from git diff output, git log output (commit messages), external issue tracker content via workflows in docs/agents/issue-tracker.md, and specification files in docs/, specs/, or .scratch/ folders.\n
Boundary markers: The instructions for the sub-agents do not include specific delimiters or warnings to ignore malicious instructions that might be embedded in the code or documentation being reviewed.\n
Capability inventory: The skill performs local git operations including git diff, git log, and git rev-parse to retrieve repository data.\n
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the repository or issue trackers before it is processed by the agent.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (git) to retrieve the diffs and logs necessary for the review. While these are appropriate for the tool's function, they represent an interaction with the underlying system that requires valid input to avoid unexpected behavior.

code-review