web3-methodology-research

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). This document is dual-use security research: it contains concrete exploit patterns, PoC code, and grep/scan commands that could enable attacks but are presented in an audit/defensive context, so it's suspicious without clearly malicious intent.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains explicit runtime fetch-and-execute instructions (e.g., "git clone https://github.com/SunWeb3Sec/DeFiHackLabs" and "go install github.com/crytic/medusa@latest") which download remote code that the workflow then runs (forge tests, medusa fuzzing), so those repository URLs are runtime dependencies that can execute remote code.