prompt-contamination-guard
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a standard prompt-engineering tool that operates entirely within the scope of text transformation. It does not contain any malicious logic, hidden commands, or unauthorized system access patterns.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided generation prompts, which serves as a data ingestion point.
- Ingestion points: Untrusted prompt text is ingested for audit and cleaning as described in the 'Workflow' and 'Contamination Classes' sections of SKILL.md.
- Boundary markers: The skill does not define specific technical delimiters for input data, though the 'Output Contract' defines a structured schema for analysis results.
- Capability inventory: The skill's functionality is limited to text auditing and rewriting; it lacks tools or scripts with file system, network, or command execution permissions.
- Sanitization: The primary purpose of the skill is to identify and remove 'contamination' (unwanted associations and instructions) from prompts, acting as a semantic sanitizer.
Audit Metadata