popey/Tessl Skill Eval
Run Tessl evals against tiles with scenarios when SKILL.md files change in a PR
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| enabled | Enable eval runs. Set to false to disable entirely. | — | true |
| skip-label | PR label that skips eval even when enabled. Set empty to disable. | — | skip-eval |
| path | Root path to search for SKILL.md files | — | . |
| comment | Whether to post results as a PR comment | — | true |
| eval-workspace | Tessl workspace name. Optional when tiles have workspace set in tile.json. | — | "" |
| eval-agent | Agent:model pair for evals (e.g. claude:claude-sonnet-4-6) | — | claude:claude-sonnet-4-6 |
| eval-timeout | Max minutes to wait for each eval run to complete | — | 45 |
| eval-fail-on-regression | Fail the check if any scenario scores worse with context than baseline | — | true |
| eval-generate-scenarios | Generate fresh scenarios before running evals | — | false |
| eval-scenario-count | Number of scenarios to generate per tile | — | 3 |
| eval-commit-scenarios | Commit generated scenarios back to the PR branch (requires contents: write) | — | false |
| tessl-token | Tessl API token for authentication. Pass via secrets. | yes | — |
Outputs
no outputs