evidence-analysis

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates within its intended scope for experiment analysis and release management. It utilizes authorized internal commands to read and update project state.
  • [DATA_EXFILTRATION]: The skill references official FeatBit documentation at docs.featbit.co. This is a vendor-owned resource used for legitimate reference in the tool adapter (tool-featbit-abtesting.md).
  • [PROMPT_INJECTION]: The skill ingests data from external database fields, which represents a potential surface for indirect prompt injection. However, the risk is low as the data is used for structured decision reporting.
  • Ingestion points: Reads primaryMetric, guardrails, and hypothesis from a database via the project-sync skill (SKILL.md).
  • Boundary markers: No explicit delimiters are defined for the interpolation of project data fields.
  • Capability inventory: The skill can modify database records using update-state, set-stage, and upsert-experiment commands (SKILL.md).
  • Sanitization: No validation or sanitization of the input strings is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 06:46 AM
Security Audit — agent-trust-hub — evidence-analysis