cognitive-bias-detection
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions and structural templates for cognitive bias analysis. It does not include any scripts, binaries, or configuration that could lead to technical security vulnerabilities.
- [NO_CODE]: There are no code files or executable segments within the skill. All logic is handled through the agent's prompt processing.
- [DATA_EXPOSURE]: No sensitive file paths, environment variables, or hardcoded credentials were detected. The skill operates solely on the context provided in the conversation.
- [REMOTE_CODE_EXECUTION]: No remote download patterns or dynamic execution mechanisms (like eval or subprocess calls) are present.
- [PROMPT_INJECTION]: The instructions are focused on self-correction and bias detection. There are no attempts to bypass safety filters or override system-level constraints.
- [DATA_EXFILTRATION]: No network tools or external domains are referenced. The skill remains contained within the local session.
Audit Metadata