claude-certification

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). These are a personal GitHub repository and its raw GitHub content URL that the skill explicitly instructs users to fetch and run code from—running code pulled from an untrusted/unknown repo can distribute malware.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly falls back at runtime to reading files from https://raw.githubusercontent.com/rohitg00/ai-engineering-from-scratch/main/, which can supply track JSON, lesson text, and executable Python (e.g., code/main.py) that the agent ingests and may execute, thereby directly controlling prompts and runtime behavior.