feat(orchestration-v2): populate subagent observability from providers (2/5)

[shivamhwp]

Second of five stacked PRs replacing #4551. Stacked on #4779 — review that first; this diff is against it, not the base branch.

PR 1 added the schema, migration, and storage with adapters on neutral defaults. This one makes the adapters actually populate it, across Codex, Claude, Cursor, OpenCode and ACP: role provenance, token usage, recent activity, and per-activation records.

Two usage strategies, because providers differ

Codex reports cumulative usage per thread, so those snapshots merge by maximum — a duplicate or late out-of-order frame then leaves the total unchanged rather than inflating it. Claude reports per activation, so lifetime usage accumulates the delta since the previous snapshot. Using either strategy for the other provider double-counts or freezes the number, so they are separate functions with their own tests.

A subagent identity is reusable, so finishing an activation leaves it at idle rather than completed. The timeline row for that activation is still recorded as completed, via PR 1's status mapping.

Three activation-lifecycle bugs found in review

All in the Claude adapter, all fixed here:

• A raced resume silently lost its tokens. When a resume arrives past settle, the wake buffer pre-opens the registry entry — which erased the terminal status the reopen check reads. The resume then reused the finished activation instead of starting a new one: its ordinal stopped matching reality, and its usage was measured against the previous activation's totals, clamping the delta to zero. The pre-open now records that it happened, and the reopen check honours it.
• A replayed waiting frame could resurrect a settled subagent. Only pending/running were guarded.
• Duplicate terminal frames re-stamped completedAt, drifting the recorded completion time. Codex already preserved it; Claude now matches.

Explicitly not here

Reuse across runs. A subagent row still carries the run it was spawned under, so a later run's update can still be dropped by routing — visible in the subagent_continue fixture. That is PR 3, deliberately kept separate because it is the subtle part.

Testing

Typecheck and lint clean. Server 1,682 passing; whole-repo run 5,460 passing (the only failures are pre-existing apps/desktop collection errors from a broken local Electron install, untouched by this PR).

Reviewed by an independent agent pass before opening; the three lifecycle bugs above came out of it and are fixed in this commit.

Note

Populate subagent observability with activation lifecycle events and usage tracking across providers

• Adds subagent_activation.updated event emission to all provider adapters (ACP, Claude, Codex, Cursor, OpenCode) so each activation has its own lifecycle record with status, timing, usage, and ordinal.
• Introduces mergeCumulativeSubagentUsage and accumulateCumulativeSubagentUsage in SubagentObservability.ts to correctly compute lifetime token/tool-use totals across multiple activations without double-counting.
• Adds appendSubagentActivity to track human-readable activity summaries on subagents, bounded to 6 entries with adjacent-duplicate deduplication.
• Adds providerSubagentRole so provider-supplied role names are tagged with source: 'provider' rather than the app default.
• Subagent status after completion is now 'idle' rather than 'completed'; currentActivationId is cleared on settlement and activationCount increments only on new activations.
• The Claude adapter gains workflow coordinator/worker subagent projection from task_progress frames, including phases, membership, attempt tracking, and resumePending flag to correctly separate activations across wake-buffer replays.
• Behavioral Change: new subagent tasks created by the Codex adapter now start in 'pending' status with activationCount: 0 instead of 'running' with activationCount: 1.

📊 Macroscope summarized 2565ea0. 9 files reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: a90349b7-5ead-47ed-9033-d9e9fec92f6e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch subagent-obs/02-adapters

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}