The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and review code diffs and file contents, which are untrusted external inputs. This introduces a surface for indirect prompt injection where instructions hidden in code comments or metadata could attempt to influence the agent's behavior.
Ingestion points: The agent reads external content via git diff and file system access to review changed code.
Boundary markers: The instructions lack specific delimiters or instructions for the agent to disregard natural language commands found within the code under review.
Capability inventory: The skill allows the agent to execute shell commands such as git, ruff, and local project scripts like ./scripts/hooks/lint.sh.
Sanitization: There are no explicit sanitization steps for processing untrusted code strings before they are analyzed by the LLM.
[COMMAND_EXECUTION]: The skill instructs the agent to execute standard development tools and local scripts to perform its review tasks.
Evidence: Execution of git log, git diff, and ./scripts/hooks/lint.sh are requested as part of the review steps. These are standard operations for a development-focused agent.

code-review