skills/reforma-ai/briefback/handoff/Gen Agent Trust Hub

handoff

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill performs local file system operations within a designated .agentflow directory to persist project state between sessions. It does not perform network operations or request elevated privileges.- [SAFE]: The instructions include a clear security requirement to redact secrets and avoid including large spec bodies or diffs, favoring file paths instead.- [INDIRECT_PROMPT_INJECTION]: The skill creates a handoff file that carries instructions into future agent sessions. This is the intended behavior for context persistence, though it creates a surface where past instructions influence future sessions.
  • Ingestion points: Reads project plans and implementation details from the current session context (SKILL.md).
  • Boundary markers: The generated file template uses Markdown headers and instructional blockquotes to bound the next agent's actions.
  • Capability inventory: Limited to writing documentation files to the local repository.
  • Sanitization: Explicitly instructs the agent to redact secrets before writing the handoff file.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 07:57 PM