headless-ghidra-frida-runtime-injection

SUSPICIOUS: the skill is internally coherent for headless reverse-engineering/runtime capture and shows no credential harvesting, exfiltration endpoint, or dubious installer behavior in the provided text. But it grants an AI agent offensive-capable instrumentation planning for Frida/Ghidra, and referenced local scripts/sibling skills were not included, so overall risk is medium despite low evidence of malicious intent.