task-executor

SUSPICIOUS-leaning but not malicious: the skill’s capabilities mostly fit its stated purpose, yet it has a broad orchestration footprint across trackers, repo writes, commits, external implementation engines, and other skills. The main concerns are transitive trust and indirect prompt-injection risk from ticket/spec content flowing into code-writing engines, plus autonomous ticket/comment/commit actions. No clear credential theft, covert behavior, malicious endpoint, or deceptive install path is present in the provided content.