authoring-signals-scouts

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill docs describe powerful, automatically-dispatched LLM-driven agents that run with a TRUSTED network, can clone/run external code, write findings/reports (including opening PRs), and are discovered/seeded automatically — a capability combination that readily enables data exfiltration, remote-code execution and supply-chain propagation if a malicious scout or a poisoned canonical skill is introduced.