handoff

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to write a document to the host's temporary directory, requiring file system operations that occur outside the active workspace.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface (Category 8).
  • Ingestion points: Untrusted conversation history and user-supplied arguments are ingested to create the summary.
  • Boundary markers: No delimiters or markers are specified to isolate the summarized content from the structure of the handoff document.
  • Capability inventory: The skill uses file-writing tools to save output to a system-level temporary directory.
  • Sanitization: The instructions explicitly require the agent to redact credentials and PII, providing a mitigation against accidental sensitive data exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 06:43 AM
Security Audit — agent-trust-hub — handoff