Skip to content

v0.36.3

Choose a tag to compare

View all tags
@github-actions github-actions released this 24 Oct 06:35
v0.36.3
e2e8ce4
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Fixes

  • Deploy to Cloud works for a setup's second destination. Since 0.9.0, deploying a setup that already had a copy in one HarnessRouter Cloud workspace to another workspace or organization always failed with "cloud answered 404: No harness with that id." A setup's first deploy was not affected. A copy is now created the way the protocol creates a harness (POST /v1/harnesses) and replaced with PUT. A copy deleted on Cloud is created again by the next deploy. (#581)
  • A codex command's output is stored whole. Output a command printed before a pause (for example echo started; sleep 3; echo done) was missing from the task's record, though the model saw it. This is a codex bug (openai/codex#28207). The output now comes from codex's own record of what the model was given, including codex's own cut of a very long output. (#580)
  • A shared link answers 503 when its Task cannot be read. If the shared Task itself could not be read for a moment, the share API answered 500. It now answers 503 with "could not be loaded just now", as it does when the link itself cannot be looked up. The page already offered Try again. (#577)
  • Deploy to Cloud stores the destination HarnessRouter Cloud names for the key. This instance no longer fills in "Default Workspace" itself. (#581)

Known issue

  • If Cloud creates a setup's copy but its answer never reaches this instance (for example a dropped connection), the next deploy creates a second copy. The first one can be seen and deleted in Cloud.

API

  • A function_call_output item has status: "incomplete" when the tool's output could not be read in full, instead of "completed". For codex, that happens when its record of the command never arrived, and always in codex's exec mode (HARNESS_CODEX_APPSERVER=0). This stays within UHP 2026-10-19: status is a string field, and "incomplete" is the value a response already uses for work that did not finish.

Checked on the test VM:

  • codex tasks: one command, two commands in a task, a 25 s command, a command polled until it exits, and a 169 KB output, each stored as the model saw it;
  • Deploy to Cloud in a QA organization: a fresh sign-in landing in "Default Workspace" as Cloud named it; the setup that had failed deploying, then deploying again to the same copy; a new setup's first deploy; and a setup uploaded by an older version, replacing its copy without making a second;
  • New setup with a first task, a share link opened signed out, API keys, workspaces, add-ons and log out.

Image: harnessrouter/harnessrouter:0.36.3 (also latest), linux/amd64 and linux/arm64, sha256:980a6ff5199f5d8e80a85c190ffe6112c790d9f16e86c02a36edcd71df0a3685.

Full Changelog: v0.36.2...v0.36.3

Assets 2
Loading