db-whisperer

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements a comprehensive multi-tiered safety framework (Red, Yellow, Green) that explicitly identifies and warns against high-risk operations such as irreversible migrations, destructive production changes, or primary key alterations.
  • [PROMPT_INJECTION]: The skill identifies a potential indirect prompt injection surface as it is designed to ingest and analyze external data sources.
  • Ingestion points: The skill processes application source code (ORM), database schema definitions, and query logs (SKILL.md).
  • Boundary markers: While the skill uses markdown code blocks for data organization, it does not explicitly instruct the agent to disregard potential natural language instructions embedded within logs or code files.
  • Capability inventory: The skill has the capability to generate and suggest SQL commands for execution, though it emphasizes the use of 'EXPLAIN' and local staging environments.
  • Sanitization: The skill incorporates best practices such as recommending 'EXPLAIN' before 'ANALYZE' and requiring rollback paths for all migrations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 01:55 AM
Security Audit — agent-trust-hub — db-whisperer