The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts, specifically scripts/security-review-scan.py and hooks/security-review-hook.py, to perform deterministic vulnerability scanning on git changes. These scripts are intended to be part of the skill's local deployment environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code changes from git diffs, which serves as a potential attack surface for indirect prompt injection if a malicious diff contains instructions designed to mislead the agent's review verdict.\n
Ingestion points: Git diff output is captured in SKILL.md via git diff --name-only and related commands in Phase 1.\n
Boundary markers: No specific delimiters or boundary warnings are used when interpolating the untrusted diff content into the agent's prompt context.\n
Capability inventory: The skill has access to Bash, Read, Grep, Glob, Task, and Agent tools, enabling it to execute scripts and manipulate the file system.\n
Sanitization: The skill does not perform explicit sanitization or filtering of the diff content before it is analyzed by the LLM.

security-review