ai-slop-cleaner
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes instructions aimed at aligning the AI's behavior with a specific persona ("GPT-5.5 Guidance Alignment"), emphasizing concise output and autonomous progression. These meta-instructions are used to govern the model's interaction style during the cleanup process.
- [COMMAND_EXECUTION]: The procedure involves executing various diagnostic and verification tools, including regression tests, linters, and static security scanners. These operations utilize local system resources to ensure code quality and behavioral consistency.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core function is to analyze and refactor potentially untrusted AI-generated code.
- Ingestion points: Files within a specified scope or those identified in a Ralph workflow (SKILL.md).
- Boundary markers: Absent. The skill does not provide delimiters or instructions to ignore embedded commands within the code being refactored.
- Capability inventory: The agent can read and write files and execute system commands for testing and linting (SKILL.md).
- Sanitization: Absent. Code is processed directly without escaping mechanisms to prevent instruction execution.
Audit Metadata