askme

Pass

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious behavior or security vulnerabilities were detected in the skill instructions. The content focuses entirely on improving the agent's reasoning process through hypothesis sampling and actor-critic analysis.
  • [PROMPT_INJECTION]: While the skill uses strong directives to guide agent behavior (e.g., 'must proactively and critically execute'), these instructions are intended to define a task-specific reasoning strategy rather than bypass safety filters or override the underlying system prompt.
  • [DATA_EXFILTRATION]: There are no commands, network requests, or file system access patterns that suggest any attempt to access or exfiltrate sensitive user data.
  • [REMOTE_CODE_EXECUTION]: The skill does not include any external dependencies, remote script fetching, or dynamic code execution mechanisms.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 5, 2026, 05:14 PM