The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external source code to perform refactoring tasks. This creates a potential surface for indirect prompt injection if the code being simplified contains malicious instructions hidden in comments or strings intended to subvert the agent's behavior.
Ingestion points: External code files provided to the agent as described in the SKILL.md process.
Boundary markers: The prompt does not specify unique delimiters or explicit instructions to ignore embedded commands within the code blocks it processes.
Capability inventory: The skill is integrated into workflows (Fixer/Oracle) that typically possess file read and write capabilities, which could be targeted by an injection attack.
Sanitization: No explicit sanitization or input validation logic is present in the prompt to filter malicious content from the code being processed.
[SAFE]: The skill includes a reference to an external GitHub repository for attribution and documentation purposes. This reference is to a well-known developer's public repository and does not involve any automated code execution or suspicious download patterns.

simplify