contract-decode

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

• Third-party content exposure detected (medium risk: 0.30). Skill.md takes user-pasted hex, extracts selector/address/chainId, then performs external reads for that data via Step 3 (Sourcify/Etherscan/4byte.directory) which ingest untrusted network responses at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime HTTP requests to external services (e.g., https://sourcify.dev/server/v2/contract/{chainId}/{address}, https://api.etherscan.io/v2/api?..., and https://www.4byte.directory/api/v1/signatures/...) to fetch ABIs/signatures that are required to decode data and thus directly influence agent behavior and outputs.