Cloudskill
Govern the AI skills your team depends on
177 followers
Govern the AI skills your team depends on
177 followers
Bring order to the AI agent skills your team depends on. Cloudskill turns scattered skill files into a managed catalogue in seconds, complete with version control, per-person access policies, and a full audit log. When skills are created or updated, every change is reviewed, approved, and tracked.
Cloudskill
@tom_palmer_ux I'm thinking about the workflow, if team members can add anything to what is approved, how are versions controlled? Or do you only assign certain team members to be contributors?
Cloudskill
@sasha_mcclendon great question, thank you! Once a skill is approved any changes are a new skill that goes through the same process. So every change creates a new skill rather than modifying the existing skill.
This is a sharp problem to take seriously. Skills are starting to look less like docs and more like production dependencies.
One question I would be curious about: do you plan to connect skill approval/versioning with runtime receipts later? For example, when an agent takes a customer-visible or state-changing action, being able to answer which approved skill version shaped the run, which tools it used, and what changed would make the governance story much stronger.
Cloudskill
@blah_mad That's a great observation and question, thank you!
To answer your question - yes, I am currently developing in that direction and should have an update soon.
@tom_palmer_ux Nice. That is probably the useful jump: not just a skill was approved, but this run used v17, called these tools, changed this state, and left a receipt the team can inspect later.
Cloudskill
@blah_mad Yep agreed! Any other feature thoughts most welcome!
@tom_palmer_ux I would probably think about drift next: when a skill changes, show which agents or workflows are now affected, and maybe block high-risk runs until the new version is approved for that use case.