autonomous-agent-patterns

SUSPICIOUS: the skill is mostly benign documentation, but it teaches high-risk autonomous-agent patterns that combine untrusted web content, prompt injection exposure, code generation, and shell execution examples. No clear malicious intent or hostile data routing is present, but several example patterns are security-sensitive and could be unsafe if implemented as written.