testing-handbook-skills

Coherent and publisher-consistent documentation skill, but it enables AI-agent-driven security testing and fuzzing workflows, which is a high-risk capability class even without obvious malicious behavior. No direct credential harvesting, exfiltration path, or suspicious installer is present in this fragment.

The skill is internally consistent and appears legitimate: it teaches use of the official LibAFL ecosystem with mostly verifiable install sources and no evident credential harvesting or exfiltration. However, it materially increases risk because it gives an AI agent offensive fuzzing capability and includes remote installer execution patterns, so it should be treated as high-risk but not malicious.