confidence-honesty
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill provides a structured methodology for agent reasoning. It does not include any executable code, scripts, or binary files.
- [PROMPT_INJECTION]: No prompt injection or safety bypass patterns were detected. The instructions focus on calibrating the AI's confidence and do not attempt to override system safety guardrails.
- [DATA_EXFILTRATION]: The skill contains no instructions for accessing sensitive files (like credentials or SSH keys) or performing unauthorized network requests.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to 'gather more evidence yourself' by searching codebases or fetching files. This is a standard functional surface for an investigative agent and no malicious processing logic is introduced. Ingestion points: references to 'Search codebase' and 'Fetch a file' in SKILL.md. Boundary markers: None. Capability inventory: generic file and search tools. Sanitization: None.
- [OBFUSCATION]: No obfuscation techniques, such as Base64 encoding of commands, zero-width characters, or homoglyph attacks, were detected in the text.
Audit Metadata