The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection risk surface identified. The skill's primary function is to interpret and modify code based on its content, which includes reading potentially malicious instructions embedded in comments or strings.
Ingestion points: The skill ingests untrusted data from the repository via the Read tool and git diff commands executed through Bash in the 'Inspect the diff' step.
Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard natural language instructions found within the code being analyzed.
Capability inventory: The agent has access to Write, Edit, and Bash tools, allowing it to modify files and execute system commands in response to ingested content.
Sanitization: No sanitization or validation of the ingested code is performed to ensure the agent does not execute instructions hidden in the 'slop' it is tasked to remove.

deslop