CUA-frontier Loading Submit a task

Team task portal

Task scenarios at the frontier of computer use

Each area studies something current computer-use benchmarks cannot see: games played on their own clock and next to people, what work done on a screen does in the physical world, chip designs whose checks can be wrong too, and processes whose final state hides the mistake. Two more cut across domains: worlds that do not wait for the agent, and work shared between people and their own agents.

Each area holds several projects, each a direction with room for many kinds of task. Contributors' agents build the tasks, each with a verifier of its own, and people review them.

Browse the areas Submit a task with your agent Board
Tasks
0
Contributors
0
Projects covered
0/15
Merged
0

Areas and projects

Each area holds several projects, each a direction with room for many kinds of task. Select one to read it, or propose a new project in any area.

Games0 tasks
Game development loop0 Start here Live play with people0 Main bet Unknown rules0 Training environment + Propose a projectin Games
Physical AIPriority0 tasks
Design, build, test0 Start here Lab experiments0 Second stage Machine programming0 New Teleoperation0 Main bet Simulated worlds0 Start here + Propose a projectin Physical AI
Chip & EDAOptional0 tasks
Design and verification0 Open Physical implementation0 Open + Propose a projectin Chip & EDA
Workflows0 tasks
Cross-system workflows0 Training environment + Propose a projectin Workflows
Real timePriority0 tasks
Real-time decisions0 Start here + Propose a projectin Real time
CollaborationPriority0 tasks
Shared game development0 Start here Asynchronous integration0 Second stage Shared 3D design0 Second domain + Propose a projectin Collaboration
Proposed In review Validated Merged

From claim to merge

Each stage maps to something on GitHub, and the board reads it from there.

  1. Proposed

    Claim a scenario (optional)

    Say which scenario your agent is generating, so others do not duplicate it. A draft pull request counts too.

    issue · label scenario
  2. In review

    Open a pull request

    Your agent builds the task in tasks/<project>/<task-id>/, with a verifier of its own that it calibrates and has a red-team agent test. Run the frontier models on your machine, read the traces, have a separate agent audit it, and commit the records. Task check comments on anything missing, and the board shows its protocol tags.

    pull request · reviews
  3. Validated

    People review it

    A reviewer other than the submitter approves it, reading the local check report and rerunning it on their machine if needed. A maintainer adds validated.

    approval · label validated
  4. Merged

    Discuss failures, merge

    Reviewers and the submitter go through how the frontier agents failed, record it in the pull request, and merge.

    merged into main

What every task needs, and how review works →