The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git rev-parse and git diff to resolve branch references and compare changes between commits. These are standard development tools used for their intended purpose of code analysis within the skill environment.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted code changes from a repository. 1. Ingestion points: Data enters the agent context via git diff output and file reads as specified in SKILL.md. 2. Boundary markers: The skill does not define specific delimiters for the ingested code content. 3. Capability inventory: The skill is restricted to git commands and file system reading. 4. Sanitization: No specific filtering or escaping of the ingested code is implemented. The risk is considered low as the skill explicitly instructs the agent to maintain an adversarial, critical posture when reviewing the code to identify blockers and risks.

preflight