output-sanitizer
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or behaviors were detected; the skill acts as a defensive security utility.
- [NO_CODE]: The skill contains only instructional markdown and lacks executable code or dependencies.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted agent output.
- Ingestion points: Agent responses processed by the module (SKILL.md).
- Boundary markers: No explicit data delimiters are defined in the instructions.
- Capability inventory: Permission 'fileRead' is defined in metadata but is not utilized in the skill's logic.
- Sanitization: Implements regex-based redaction for various secret types and PII.
Audit Metadata