handoff-invoke

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes task summaries and context from local files in the .handoffs/ directory, creating a surface for indirect prompt injection.
  • Ingestion points: Content is read from markdown files located in the .handoffs/ directory during the resolution and briefing steps (SKILL.md).
  • Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" markers when reading the content of handoff files.
  • Capability inventory: The agent uses file system read capabilities and project state verification (checking for files and git branches) based on the content of the handoff files.
  • Sanitization: There is no explicit sanitization for malicious instructions, although the skill includes a security directive to avoid repeating secrets or credentials in the chat output.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 08:42 PM
Security Audit — agent-trust-hub — handoff-invoke