vet-flow-guide

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted external sources, creating a surface for indirect prompt injection. * Ingestion points: The skill reads JSON run files (.skills//flow-guide.json), markdown specifications (requirements.md, design.md, tasks.md), and arbitrary product source code files during its mapping phase. * Boundary markers: The instructions mandate isolation through subagents or an 'AUTHORING CLOSED' session marker, along with a set of 'Iron Law' rules and 'Read-only rules' to prevent context bleeding and unintended mutations. * Capability inventory: The agent can write report files to the filesystem (.skills//vet-flow-guide.md) and dispatch subagents with specific briefs. * Sanitization: The skill instructions do not specify any sanitization or validation logic for the content of the files it reads before processing them or including them in findings.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 11:40 AM
Security Audit — agent-trust-hub — vet-flow-guide