write-techspec

SUSPICIOUS. The skill's core behavior is coherent for tech-spec authoring and its flagged command-injection/autonomy findings look mostly benign or false-positive, but it depends on an externally supplied `roundfix` CLI whose provenance could not be verified. That makes the footprint higher risk than the documentation purpose alone would suggest, though there is no strong evidence of credential theft or malicious exfiltration in the skill text itself.