caido-mode

The skill is internally consistent and uses Caido-aligned dependencies and data flows, so it does not look malicious. However, it equips an AI agent with substantial offensive web-testing capability and persistent credentialed access, making it a high-risk security tool despite coherent purpose alignment.