The Agent Skills Directory

[SAFE]: The skill performs string manipulation to repeat input prompts for optimization purposes. It contains no network calls, file system access, hardcoded credentials, or persistence mechanisms.- [PROMPT_INJECTION]: Indirect prompt injection surface analysis: Ingestion point is the 'prompt' parameter in the 'transform' method within 'prompt_repetition_transformer.py'. Boundary markers are present as internal tracking tags but do not isolate user-provided content. Capability inventory confirms the absence of hazardous operations (e.g., subprocess execution, network requests, or file writes). Sanitization is not performed as the tool is designed to repeat input text exactly. The risk is negligible because the skill possesses no operational capabilities to act upon injected instructions.

prompt-repetition