etherscan-contract-review

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path ingests outsider-authored free text from Etherscan: it uses the authenticated CLI/API to retrieve verified contract source/ABI/metadata (including comments/strings) for the address the user provides, and then analyzes that retrieved bundle.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent at runtime to read live external documentation/repositories which are then used to construct installation or retrieval commands (e.g., https://github.com/etherscan/etherscan-cli/, https://docs.etherscan.io/endpoint-overview, https://docs.etherscan.io/common-error-messages), so those URLs are runtime dependencies that can directly control prompts/instructions.