Jamf's enterprise AI team built a guided performance self-evaluation tool as a Claude Cowork skill in under 45 minutes, replacing a spreadsheet covering seven competency areas.
- First reported
- claude.com · 1 Jun 2026
- Added
- 13 Oct 2026
- Source checked
- 13 Oct 2026
What they did
The skill asks for the person's name, level and role, builds a personalised competency checklist, asks about specific behaviours and pushes for concrete examples, and can pull recently closed Jira tickets when someone gets stuck. It outputs a file to paste back into the original spreadsheet plus a summary with a level assessment and development plan. The team also built a vendor review skill that researches a software vendor, drafts answers to a 12-question review and walks the reviewer through each one.
What happened
The builder says a comparable build at a previous employer took a full team three months, and that staff across departments now make bespoke dashboards in minutes without a BI tool or engineer.
Keep in mind
The builder is a software engineer, a self-review session still takes 60 to 90 minutes, and the team stresses that people still make the decisions.
Try it with your agent
Create a skill that walks an employee through a self-review using our competency framework in <your framework file>. Ask for their level and role, show a checklist of the competencies that apply, and ask one question at a time, pushing for a concrete example each time. At the end, write a summary with strengths, gaps and a three-item development plan, and export answers in the column layout of <your review spreadsheet>.
Replace the <angle brackets> with your own details.
Original source
claude.comVendor case studyPublished by the company that makes the agent. Read it as a claim.
Tools for this kind of work
- RepoManage your Supabase database and projects from your agent
- RepoManage GitHub issues, pull requests and repos from your agent
