closing-checklist
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, potentially untrusted sources including legal contracts, data room files, and external communications via a monitoring agent.
- Ingestion points: The skill reads "Equity Purchase Agreements" and "material contracts" to extract closing conditions (Mode 1). It also receives findings from other skills like
diligence-issue-extractionandmaterial-contract-schedule(Mode 2), and pulls status updates from email or Lark (Feishu) via a monitoring agent integration. - Boundary markers: The instructions lack explicit boundary markers or delimiters to differentiate between the skill's operational instructions and the potentially instructional content found within the analyzed legal documents.
- Capability inventory: The skill writes and updates
closing-checklist.yamlfiles within the~/.claude/plugins/config/directory and generates status reports that can be pushed to deal team channels. - Sanitization: There is no evidence of sanitization or validation logic for the input text extracted from external documents or messages before it is processed or stored in the YAML checklist.
Audit Metadata