cross-agent-handoff

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a framework for agents to ingest and process handoff documents and project plan files from previous sessions.
  • Ingestion points: The skill instructs agents to read handoff documents and plan files (e.g., .claude/plans/*) as documented in the 'Recovery Protocols' section of SKILL.md.
  • Boundary markers: While the skill uses structured markdown templates, it does not prescribe explicit delimiters or instructions to treat handoff content as untrusted data.
  • Capability inventory: The skill itself does not contain subprocess calls, network operations, or privileged file access; the logic is limited to string compression and conflict detection.
  • Sanitization: The provided logic lacks mechanisms to sanitize or filter potential instructions embedded within the handoff data.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 04:22 PM