database-design
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill incorporates a strict command-execution policy (BF-COMMAND-1) that limits operations to necessary and safe actions, prioritizing read-only inspection and requiring user authorization for any privileged or destructive tasks.
- [DATA_EXFILTRATION]: The Security and redaction policy (BF-SECURITY-1) prevents the exfiltration of sensitive information by mandating the redaction of secrets in all outputs and handoffs.
- [INDIRECT_PROMPT_INJECTION]: The skill includes an Untrusted-content policy (BF-UNTRUSTED-1) that provides clear instructions to treat all repository files and external data as evidence rather than authority, specifically instructing the agent to ignore any embedded prompt-injection directives.
- [DYNAMIC_EXECUTION]: The package contains a static validation script (scripts/validate_scenarios.py) used for checking the integrity of the skill's documentation, but this script is not invoked during the agent's task execution and contains no malicious logic.
Audit Metadata