Journey: Get a skill or model measured
← UX reviews · Persona: Skill builder
Goal: a builder gets their skill, prompt, model or tool scored on an open benchmark and listed on the leaderboard. This is the Measure step, and it can lead to Release.
Done when: a leaderboard row with the builder’s solution is live, with its score, cost and credit.
Steps today (2026-11-18)
| # | Where | What the person does |
|---|---|---|
| 1 | Homepage or SUBMIT.md | Reads how to submit. Picks a task and its benchmark. |
| 2 | Their own machine | Runs the solution on the test data and scores it with the task’s scorer, at least 3 times if it is not deterministic. |
| 3 | buddhist-benchmarks | Opens a pull request with result.json, outputs, scores.json and a skills.json line. Keeps a provenance record in their own repo. |
| 4 | Maintainer | Checks and merges. CI rebuilds the leaderboard. |
| 5 | Release (if it leads) | The skill is promoted to buddhist-skills, and the new leader is announced. |
Planned: submit by pasting a repo link, scoring by a GitHub Action, and maintainer approval (team call 2026-11-17). Kalden’s leaderboard app will change steps 1 to 4.
Pages this journey touches
Homepage submit box, SUBMIT.md, docs/results.md, task TASK.md pages, the leaderboard site and app, the provenance handbook page, the workshop template repo.

BuddhistAI Knowledge base