gpt-image2-ratos

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). While the OpenAI developer and status pages and localhost endpoints are legitimate, the workflow depends on running an unvetted npm package (npx openai-oauth) and small GitHub repos (EvanZhouDev, duduesh) that could contain malicious code or exfiltrate local auth tokens—so the aggregate set poses a notable risk if you execute the referenced packages/scripts without auditing them.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill invokes "npx -y openai-oauth" at runtime, which fetches and executes remote code (the openai-oauth proxy) that the skill requires to operate — see https://github.com/EvanZhouDev/openai-oauth.