The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the repository's source code and documentation, which constitutes an ingestion point for indirect prompt injection. • Ingestion points: The skill reads files and git diffs within the local repository, including comments and docstrings. • Boundary markers: No explicit delimiters or safety instructions are used to isolate untrusted data; the skill relies on logical 'Gates' to constrain behavior. • Capability inventory: The skill can perform file writes to modify the code and execute system commands for testing and type-checking. • Sanitization: No sanitization of the input text is described before it is processed by the agent.
[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands to verify the integrity of the cleanup. The 'Output' section instructs the agent to run 'typecheck' and 'tests' on the local system. This execution occurs after the agent has modified the code, creating a path for potentially malicious code to be executed during the validation phase.

hl-deslop