Skip to content

Damson 0.7.0

Choose a tag to compare

View all tags
@github-actions github-actions released this 10 Sep 02:58
v0.7.0
181c6e2
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

What's new

Managed agent workflows

Build projects through dependent agent tasks with damson-crew workflow run, status, and validate. Finite commands run in real Damson panes, with bounded parallelism, shared-resource exclusion, validation commands, required outputs, timeouts, and retries. Execution records and logs let the coordinator resume without duplicating live or completed attempts. If a worker wrapper disappears, surviving commands are identified and the failure is reported without automatically repeating their work.

The orchestration plugin is now version 0.3.0 and guides task decomposition, parallel implementation, integration, recovery, and final artifact checks. Existing interactive run and watch commands remain available.

More reliable agent management

  • Prevent duplicate/cross-group task launches and restore the correct task after a restart.
  • Reconnect watchers across app restarts and focus the exact waiting pane across windows and splits.
  • Protect shared, dirty, externally created, and in-use worktrees during cleanup; allow safe cleanup retries.
  • Fix worktree path handling, rapid session-status changes, and Claude argument/trust handling.

Terminal and installer improvements

  • Drag tabs between windows.
  • Close tabs with the middle mouse button and switch tabs using the mouse wheel over the tab bar.
  • Refreshed DMG installer window.

Using managed workflows

The updated damson-cli and damson-crew binaries are bundled with the app. Update the damson-orchestration Claude Code plugin to 0.3.0 as well; the app updater does not update that plugin automatically.

See the workflow guide and acceptance audit.

Validation

273 Swift tests (one existing skip, no failures), 17 CLI integration tests, and isolated actual-app checks passed. Actual Claude workers produced a playable Snake game, followed by interruption recovery, 53 game tests, and desktop/mobile Chromium play checks. A separate natural-language planning exercise produced and validated a dependency graph.

Full changelog: v0.6.1...v0.7.0

Assets 4
Loading