photo-agents-autonomous-llm

SUSPICIOUS. The skill is internally coherent as an autonomous agent framework, but its footprint is unusually broad and high-impact: it combines multiple credentials, screen/browser access, arbitrary file I/O, code execution, long-running autonomy, and external service calls. Installation source is not overtly malicious, yet publisher/release trust is only partially verified. Main concern is not clear credential theft but disproportionate operational risk and prompt-injection exposure for an AI agent skill.