review-for-readers

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided documents by passing them to a sub-agent. If these documents contain malicious instructions, they could potentially override the sub-agent's behavior during the review process.
  • Ingestion points: The skill accepts the full text or file path of a document as input for review in the '読み手の準備' (Reader Preparation) section of SKILL.md.
  • Boundary markers: The skill includes an explicit instruction to tell the sub-agent 'not to open files other than the document passed', which serves as a basic boundary to limit the scope of the sub-agent's actions.
  • Capability inventory: The skill involves reading document content and coordinating with other text-processing skills like 'writing-rules' and 'yomiyasu'. No network exfiltration or sensitive file system write capabilities are directly exposed in this skill.
  • Sanitization: There is no evidence of sanitization or filtering of the input document content before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 12:51 AM