The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its design, which involves reading and interpreting untrusted data from a codebase.
Ingestion points: Phase 2 (Explore) instructs the agent to use Glob, Grep, and Read tools to investigate external files within the project codebase.
Boundary markers: The instructions do not define delimiters or provide warnings to the agent to disregard instructions found within the files it analyzes.
Capability inventory: The skill leverages standard file system tools and interacts with local environment configurations.
Sanitization: No mechanisms for escaping or validating the content of the ingested code are present, allowing potential malicious instructions in a repository to influence the agent's behavior.
[COMMAND_EXECUTION]: The exploration phase utilizes the git log command to extract context regarding recent changes in the repository. This is an expected and standard behavior for development-focused skills.

aep