Ecosystem review / Worker Coordination
Worker Coordination: review and next steps
Worker should give operators and agents a coherent view of task ownership, progress and next action. It should expose the existing coordination system rather than create another source of authority.
What the evidence supports
The retained host contract treats Worker as a facade over shared coordination. The programme pointer still lists Worker remote-request work as open. This review did not inventory every installed Worker service or private principal store.
Reevaluation: Start with installed-state and ownership evidence. A CLI command, queued request or finished session is not proof that the underlying task is accepted.
The next useful outcome
The coordinator can identify the running task, its owner, dependencies, failure, continuation and final acceptance across all three principals.
Separate workstream plan
WC-01 · Proposed / reuse existing work
Refresh the installed census
Inventory enrolled principals, Worker/Steward identities, services and supported routes on each named machine.
Acceptance: Every target is classified installed, qualified, partial or unavailable with dated evidence.
WC-02 · Proposed / reuse existing work
Complete the control surface
Close demonstrated gaps in Worker control commands, supervised mode and health reporting.
Acceptance: Commands report truthful state through the existing authority route.
WC-03 · Proposed / reuse existing work
Wire typed task continuation
Reuse the landed next-action contract for task intake, dependency routing and continuation of the same task.
Acceptance: A partial slice resumes without advancing dependents or losing cumulative limits.
WC-04 · Proposed / reuse existing work
Accept cross-host state and fencing
Qualify shared PostgreSQL coordination and account-bound cross-host ownership.
Acceptance: Conflicting writers and foreign-principal effects are refused and recoverable.
WC-05 · Proposed / reuse existing work
Complete traceability
Join plan, packet, session, worktree, branch, evidence and acceptance identities.
Acceptance: An operator can trace any reported outcome back to the task and its accepted result.
WC-06 · Proposed / reuse existing work
Run the three-principal exercise
After the narrow route passes, run one bounded unattended coordination scenario across the named principals.
Acceptance: Requests finish or stop with a truthful owned blocker; no manual state copying is required.
Dependencies and sequence
Shared coordination state, A2A completion, Steward continuation and task-to-evidence traceability.
Start with the first eligible unfinished task above, after checking its existing owner and admission. Later tasks depend on the preceding proof where applicable; this ordering does not override a current implementation plan.
Deferred work
A second scheduler, a second ownership database and broad unattended operation before the pilot route is qualified.
Evidence and updates
This article combines the supplied 8 November ecosystem review with selected current source and retained planning records. It is an editorial snapshot, not a live service census. Detailed revision bindings and task authority remain in Workspace Docs and private execution records. Publish subsequent accepted changes as linked engineering or release notes rather than silently rewriting this baseline.
Shared roadmap · All Worker Coordination articles · Explore all areas
