context-handoff

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill creates a persistence mechanism for session context by writing to docs/superpowers/handoff.md. This file is designed to be read by the agent at the start of new sessions, creating a surface where instructions embedded in the file can influence agent behavior.
  • Ingestion points: The file docs/superpowers/handoff.md is automatically read by the agent during session initialization (as described in the "Resume is automatic" section referencing CLAUDE.md).
  • Boundary markers: The skill uses YAML frontmatter and Markdown headers to structure the handoff data, which provides structural separation but lacks explicit instructions for the agent to ignore potentially malicious commands within the prose sections.
  • Capability inventory: The skill performs file system operations, including directory creation and writing to a fixed path within the repository.
  • Sanitization: No explicit sanitization or escaping of session data is performed before it is written to the handoff file, allowing raw session content to be persisted.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:34 AM
Security Audit — agent-trust-hub — context-handoff