Instrument · probe suiteWhat the console points at, and what it will ask.
Every run goes to the provider with the current model selection. The suite is chosen for you: seven categories, 47 probes, each written to surface a capability or a limit rather than a feature tour.
ModelDopamine Ultra 1 ・ 10M context
Platformunverified in UI8080.AI · agentic software engineering
Probe categories
Planned probes- 01
Context depthLoad a long workspace spec and check how much of it holds across a multi-step task.
0probes - 02
Tool useAsk the agent to open its editor, run a command, and report the actual output.
0probes - 03
Multi-agent handoffConfirm that Tech Lead, Frontend, Backend, DevOps, QA, and AI Engineer agents pass work along.
0probes - 04
Session memoryClose and reopen a project to see whether earlier decisions persist.
0probes - 05
Build and test outputHave the agent produce a codebase with tests, Dockerfiles, and CI workflow files.
0probes - 06
Model gatewayRead which model the gateway reports for the current selection and what context size it states.
0probes - 07
Credit wallPush a long-running task until processing stops and record exactly what the provider says, if anything.
0probes
Counts are the probes planned per category; nothing here reports a run. Results and any credit-wall evidence appear on Probe Results once a run has executed.
No comments yet. Be the first!