Fix Daytona computer-use image and add a single-task runner - #49
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #46, closes #47.
What changed:
x11vncandnovnctoDESKTOP_PACKAGESinexamples/node22-bookworm-computer-use-image.tsso Daytonasandbox.computerUse.start()can find the VNC/NoVNC bridge it needs.scripts/run-daytona-task.ts: a supported, type-checked,runSingle-based path for running one benchmark task end-to-end through a Daytona sandbox. It uses the built-in computer-use image by default, accepts--snapshot, and always awaits the benchmark grader before printing the result, so an unawaitedscore()Promise can no longer silently serialize as{}.README.md.Verification:
npm run validatepasses (22 test files, 296 tests).bountybench / lunary-0-exploitwithtransport: "daytona"andmodel: kimi-azure/kimi-k2.6on the Hetzner server is the acceptance test for Computer-use fails to start on every Daytona sandbox: missing x11vnc/novnc packages #46/No supported path to run+score a single Daytona-transport task, forcing hand-rolled scripts that can silently drop scoring #47; non-emptyoracleScoremeans the harness reached the grader.Note
Low Risk
Developer tooling and image package additions with unit tests; no changes to auth, scoring logic, or production CLI paths beyond the new optional script.
Overview
Fixes the Node 22 bookworm computer-use image so Daytona
computerUse.start()can bring up the VNC/NoVNC stack by addingx11vncandnovncto the desktop package list, with a test that asserts both appear in the generated Dockerfile.Introduces
scripts/run-daytona-task.tsas the supported way to run one benchmark task end-to-end in a Daytona sandbox (default inline bookworm image or--snapshot). It wires anautobrincontender withtransport: "daytona"throughrunSingle, prints the full JSON result, and fails with exit code 1 when the oracle score is missing or has no signals—addressing hand-rolled scripts that could print{}from an unawaited grader. README documents usage and optional flags.Reviewed by Cursor Bugbot for commit 1bf345e. Configure here.