tao-run-inference-service

SUSPICIOUS: the skill’s stated purpose matches its main capabilities, and its secret-handling guidance is mostly appropriate, so there is no strong evidence of malware. Risk comes from transitive trust: it delegates execution to other skills, runs external container images, forwards secrets into runtime environments, and may expose inference endpoints over remote platforms; without reviewing the companion platform skills and image provenance, the overall footprint is broader than this single skill reveals.