empiricist
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely composed of markdown instructions and a YAML configuration for a persona that reviews data claims. It does not invoke any tools, scripts, or external resources.
- [REMOTE_CODE_EXECUTION]: No patterns for remote code execution, shell commands, or package installations were found.
- [DATA_EXFILTRATION]: There are no commands or instructions that involve reading sensitive data or transmitting information to external servers.
- [PROMPT_INJECTION]: The instructions set a specific tone ('ruthless Empiricist') but do not attempt to bypass safety guidelines, extract system prompts, or override core agent safety constraints.
- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted user data (experimental claims), it lacks any exploitation capabilities (e.g., file writing, network access, or command execution) that could be triggered by malicious input.
Audit Metadata