handoff

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests current session context, including linked artifacts and project history, to generate a handoff document or background agent command.
  • Ingestion points: The skill reads session history, linked specs, plans, ADRs, issues, and commits (SKILL.md).
  • Boundary markers: The template provides structure using Markdown headers but lacks explicit markers to prevent the agent from obeying instructions embedded within the session data being summarized (SKILL.md).
  • Capability inventory: The skill uses shell commands to create temporary files (mktemp) and launch new background agent processes (claude --bg) (SKILL.md).
  • Sanitization: The skill explicitly instructs the agent to redact credentials, secrets, and personal or customer data (SKILL.md).
  • [COMMAND_EXECUTION]: The skill utilizes shell commands to manage temporary files and orchestrate new agent sessions.
  • Evidence: The Procedure section specifies the use of mktemp for file creation and claude --bg for starting background agents (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 02:00 AM
Security Audit — agent-trust-hub — handoff