contract-review

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required workflow is to “upload your contract file (PDF, DOCX, or paste text)” and then the agent analyzes that contract text; that contract body is outsider-authored free text (the other party’s document) and is ingested into the LLM context for clause/risk extraction.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The README instructs fetching runtime skill content from https://raw.githubusercontent.com/claude-office-skills/all-in-one-skills/main/contract-review/SKILL.md (and references https://github.com/claude-office-skills/contract-review-skill), which would load external instructions that directly control the agent's prompts.