skills/davidlee/doctrine/handover/Gen Agent Trust Hub

handover

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data from project files to generate instructions that influence subsequent agent sessions.
  • Ingestion points: The skill reads content from multiple external files including notes.md, plan.toml, phase-NN.md, design.md, and other unspecified 'governing artifacts'.
  • Boundary markers: Absent. The instructions do not provide delimiters or guidance to the agent to ignore potentially malicious instructions embedded within the ingested artifacts.
  • Capability inventory: The skill writes to handover.md (which serves as the initial context for the next agent) and executes environment-specific commands such as /harvest, /route, and doctrine slice.
  • Sanitization: Absent. Data from the ingested files, such as 'open ids' and status updates, is interpolated directly into the continuation prompts and the handover packet without filtering or escaping.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 09:59 AM
Security Audit — agent-trust-hub — handover