claude-handoff

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the claude CLI to start background agents. The command claude --bg --name "<descriptive name>" "<handoff summary>" is constructed using LLM-generated strings based on the conversation history and user arguments.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because the summary of the current conversation becomes the initial prompt for the next agent.
  • Ingestion points: User arguments via argument-hint and the existing conversation history are used to generate the handoff summary.
  • Boundary markers: The skill uses double-quote delimiters within the shell command template to wrap the generated content.
  • Capability inventory: The skill executes the claude command-line tool via a subprocess call.
  • Sanitization: The instructions explicitly command the agent to redact sensitive information such as API keys, passwords, and PII before passing the summary to the new agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 04:57 PM
Security Audit — agent-trust-hub — claude-handoff