Who Maintains Agent Skills? A Longitudinal Study of Human-Governed, AI-Assisted Skill Maintenance

Agent skills are reusable instruction files that need updating as tools and workflows change. A study of five public repositories examined 254 substantive edits and found that every edit passed through a named human account as author or merger, while 62% included an AI co-author credit. Additions and corrections dominated the recorded maintenance. Research abstract.

The evidence describes human-governed maintenance with visible AI assistance in this selected sample. It gives researchers building automatic skill curators an observed process to compare against. A co-author credit records acknowledged assistance; it does not measure how much of an edit the AI produced.

“every substantive edit is authored or merged through a named human account, while 62% carry an AI co-author trailer, with large repository-level variation”

“public skill maintenance currently looks less like an autonomous pipeline than a human-governed, AI-assisted loop that future curators must measure against and operate within”