Template
[Agent name]
Vendor: [vendor] · Track: [Coding / General] · Plan and version tested: [plan, version] · Last tested: [date]
This is a blank template. No data has been collected for any agent.
Score /100–
Finish rate–
Median cost/run–
Avg interventions–
Per-task results
| Task | Runs passed (of 3) | Median cost | Median time | Interventions | Logs |
|---|---|---|---|---|---|
| T1 Fix a seeded bug | – | – | – | – | – |
| T2 Add a feature, open a PR | – | – | – | – | – |
| T3 Cited research brief | – | – | – | – | – |
| T4 One-page website | – | – | – | – | – |
| T5 CSV to weekly report | – | – | – | – | – |
Notes
[Where it succeeded, where it failed, notable interventions, and links to raw logs.]