hypothesis-design

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill manages project states by retrieving and updating metadata through the project-sync tool. These operations are limited to project-specific fields like goal, hypothesis, and primaryMetric, which are essential for its primary function of hypothesis design.
  • [SAFE]: Although the skill ingests data from a database (such as the project goal and intent), this ingestion surface is used purely for generating descriptive hypotheses. The skill's capabilities are restricted to project metadata updates and do not include high-risk operations like arbitrary command execution or unauthorized network communication.
  • [SAFE]: The instructional content and reference templates are pedagogical and focus on ensuring business logic is falsifiable and well-formed. No evidence of prompt injection, obfuscation, or persistence mechanisms was found.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 06:46 AM
Security Audit — agent-trust-hub — hypothesis-design