How does Endpoint AI relate to provisioning skills?

Align endpoint deployments with the same skill catalog and restrictions admins approve in Harriet, then provision approved skills to teams or users.

Provisioning skills under Endpoint AI means employees on managed or provisioned endpoints see the same Harriet capabilities your admins approved—without each user hand-installing tools. After items leave the review queue, admins provision those skills to the right teams, groups, or users.

What to align

Which skills are default vs pilot-only (including package skills submitted via Skilify).
Which MCP connectors may run on endpoints versus browser-only surfaces, and which remain in draft until reviewed.
Update windows when you promote a skill from pilot to general availability.

Suggested process

HR and IT agree the minimum skill set for the first wave (for example knowledge and tickets only).
Security reviews any external tool skills and connectors before they appear on endpoints.
Authors submit package skills through Skilify or admins create skills in Endpoint AI; reviewers use the review queue before wide rollout.
After approval, admins assign or provision skills to profiles, groups, or individuals—then verify with a small cohort before broad rollout.

Example

Week 1 endpoints get Knowledge and Tickets only. Week 4 adds Calendar after OAuth guidance is published internally. New Skilify submissions stay in review until IT attaches any needed MCP connectors and assigns the approved skill to the pilot group.

Guardrails

Skills that can write to other systems deserve the same review as integrations in Company settings.
Document rollback: how to disable a skill centrally if an incident occurs.

See Using Skilify to build skills locally, submit for review, and provision them for the full author → review → admin provisioning flow.