fix(ai-harness): a forwarded command ends with the owner's result or error by tombeckenham · Pull Request #1872 · TanStack/ai

tombeckenham · 2026-10-22T07:50:00Z

When a durable host forwards session.runCommand(...) to the host that owns the thread, the caller got completed with undefined. This was true also when the owner refused the command or the command threw. Now the forwarding host waits for the owner to settle the input, and ends the operation the same way: with the result, with the error, or with InputRejectedError.

🎯 Changes

executeCommand in packages/ai-harness/src/session/turn.ts: an offered admission now goes to a new followCommand. It reads the owner's harness.command.result event from the feed and waits on settled({ inputId }).
Three tests in tests/input-lane.test.ts: a forwarded command that returns a value, one the owner refuses (unknown_command, invalid_input), and one that throws.
One bullet in docs/harness/durable-sessions.md, and one sentence in the unreleased ai-harness-core changeset. The harness is not released yet, so no new changeset.

Fixes for tests that already failed on feat/harness-core, so CI goes green:

subagent-waits.test.ts "waits for an agent that an agent tool calls, and both go on" is skipped, with a link to ai, ai-persistence: a nested subagent that waits runs again from the start under withPersistence #1878. Since fix(ai-persistence): resume interrupts of tool-mode subagents under withPersistence #1806, a nested subagent that waits runs again from the start after the resume. A fix that copies the nested card from the client would trust the client's messages for that child, so this PR does not ship one. ai, ai-persistence: a nested subagent that waits runs again from the start under withPersistence #1878 fixes it from the store.
harness-text.test.ts: the test still expected a new session per call with no thread id. threadId is required now, so it uses one thread id per call.
conformance-skip-reporting.test.ts (from main, fix(ai-persistence): report omitted generation stores as skipped in the conformance suite #1847): on this branch memoryPersistence() has a log store, so the conformance suite needs a twin. It now uses a shared log state and a twin, as memory.conformance.test.ts does.

✅ Checklist

I have followed the steps in the Contributing guide.
I have tested code changes locally with pnpm run test:pr, or these tests do not apply to this pull request.
I fully understand the code in this pull request, including any code generated with AI assistance.
Docs: I updated docs/ for this change, or this change is not user-facing.
Changeset: I added a changeset (pnpm changeset), or this PR does not change a published package.

🚀 Release Impact

This change affects published code, and I have generated a changeset.
This change is docs/CI/dev-only (no release).

Root cause

Issue. Two durable hosts share one thread. Host A owns it and runs a turn. hostB.runCommand(...) resolves with undefined, also for an unknown command, bad input, or a command that throws.
Cause. executeCommand treated the offered admission like a duplicate. It called operation.finish('completed', undefined) at once, before the owner ran the command.
Fix. For offered, the host waits on settled({ inputId }). That rejects with InputRejectedError when the owner refuses the input. A failed or aborted settlement fails the operation. A completed settlement finishes it with the result from the owner's harness.command.result event, matched by the owner's operation id.

Possible alternatives

Receipt only. End the forwarded operation with a queued receipt and document that a forwarded command has no result. The caller still cannot tell a refused command from one that ran, so this PR did not take it.
Result in the settlement record. The owner writes the result into harness.input.settled. This changes a log record shape for one case, and the result event is already in the log.

Testing

Commands run.

Repro (agent-written, this session): npx vitest run tests/input-lane.test.ts -t "a command" in packages/ai-harness.
- On clean feat/harness-core (79a946689) plus the tests only: 3 failed. expected undefined to be 'echo hi', and promise resolved "undefined" instead of rejecting (twice).
- With the fix: 3 passed.
tests/input-lane.test.ts in full: 27 passed. The whole @tanstack/ai-harness suite passes (one test skipped, see above). ai-persistence conformance-skip-reporting: 87 passed.
pnpm test:types and pnpm test:oxlint in packages/ai-harness: pass (only old warnings). oxfmt --check: pass.
Skipped: full pnpm test:pr and E2E. CI runs them. The E2E suite has no two-host harness setup.

Manual test.

Open one thread on two hosts that share a log store (see twoStores() in tests/input-lane.test.ts).
Start a turn on host A, so A owns the thread.
On host B, call runCommand({ name: 'unknown' }). Before this PR it resolved with undefined. Now it rejects with InputRejectedError and the reason unknown_command.
On host B, run a command that returns a value. The operation resolves with that value.

How this PR makes testing easy. The three tests under a command in tests/input-lane.test.ts are the repro.

Linked issues

Closes #1796

Risk / rollback

Low. Only a command forwarded to another host changes. If the owner's result event is evicted from the feed before the input settles, the result is undefined, as before. Revert the PR to undo.

…error A command that reaches a host that does not own the thread now waits for the owner to settle it. It resolves with the owner's result, fails when the command throws, and rejects with InputRejectedError when the owner refuses it (unknown_command, invalid_input). Before, it completed with undefined. Closes #1796

coderabbitai · 2026-10-22T07:50:10Z

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

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

Source mode: Auto
Repository: TanStack/ai
Branch: feat/harness-core
File: .coderabbit.yaml
Commit: 79a9466
Layers: Repository YAML
Review profile: CHILL
Plan: Advanced
Run ID: c63811c1-c0ef-4ba2-bda2-179be780dbd5

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

Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

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

nx-cloud · 2026-10-22T07:53:01Z

View your CI Pipeline Execution ↗ for commit 7a0b833

Command	Status	Duration	Result
`nx run-many --targets=build --exclude=examples/...`	✅ Succeeded	5s	View ↗

☁️ Nx Cloud last updated this comment at 2026-10-22 08:43:36 UTC

pkg-pr-new · 2026-10-22T07:54:52Z

Open in StackBlitz

@tanstack/ai

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai@1872

@tanstack/ai-acp

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-acp@1872

@tanstack/ai-angular

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-angular@1872

@tanstack/ai-anthropic

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-anthropic@1872

@tanstack/ai-bedrock

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-bedrock@1872

@tanstack/ai-byteplus

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-byteplus@1872

@tanstack/ai-claude-code

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-claude-code@1872

@tanstack/ai-client

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-client@1872

@tanstack/ai-cloudflare

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-cloudflare@1872

@tanstack/ai-code-mode

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-code-mode@1872

@tanstack/ai-code-mode-snippets

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-code-mode-snippets@1872

@tanstack/ai-codex

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-codex@1872

@tanstack/ai-cohere

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-cohere@1872

@tanstack/ai-compaction

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-compaction@1872

@tanstack/ai-computer-local

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-computer-local@1872

@tanstack/ai-computer-playwright

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-computer-playwright@1872

@tanstack/ai-computer-sandbox

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-computer-sandbox@1872

@tanstack/ai-computer-use

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-computer-use@1872

@tanstack/ai-devtools-core

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-devtools-core@1872

@tanstack/ai-durable-stream

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-durable-stream@1872

@tanstack/ai-elevenlabs

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-elevenlabs@1872

@tanstack/ai-event-client

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-event-client@1872

@tanstack/ai-fal

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-fal@1872

@tanstack/ai-gemini

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-gemini@1872

@tanstack/ai-grok

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-grok@1872

@tanstack/ai-grok-build

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-grok-build@1872

@tanstack/ai-groq

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-groq@1872

@tanstack/ai-harness

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-harness@1872

@tanstack/ai-harness-cli

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-harness-cli@1872

@tanstack/ai-isolate-cloudflare

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-cloudflare@1872

@tanstack/ai-isolate-daytona

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-daytona@1872

@tanstack/ai-isolate-e2b

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-e2b@1872

@tanstack/ai-isolate-node

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-node@1872

@tanstack/ai-isolate-quickjs

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-quickjs@1872

@tanstack/ai-isolate-quickjs-bun

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-quickjs-bun@1872

@tanstack/ai-llmgateway

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-llmgateway@1872

@tanstack/ai-lovable

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-lovable@1872

@tanstack/ai-mcp

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-mcp@1872

@tanstack/ai-memory

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-memory@1872

@tanstack/ai-mistral

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-mistral@1872

@tanstack/ai-octane

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-octane@1872

@tanstack/ai-ollama

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-ollama@1872

@tanstack/ai-ollaya

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-ollaya@1872

@tanstack/ai-openai

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-openai@1872

@tanstack/ai-opencode

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-opencode@1872

@tanstack/ai-openrouter

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-openrouter@1872

@tanstack/ai-perplexity

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-perplexity@1872

@tanstack/ai-persistence

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-persistence@1872

@tanstack/ai-preact

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-preact@1872

@tanstack/ai-react

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-react@1872

@tanstack/ai-reactor

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-reactor@1872

@tanstack/ai-remix

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-remix@1872

@tanstack/ai-sandbox

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox@1872

@tanstack/ai-sandbox-blaxel

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-blaxel@1872

@tanstack/ai-sandbox-boxd

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-boxd@1872

@tanstack/ai-sandbox-cloudflare

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-cloudflare@1872

@tanstack/ai-sandbox-daytona

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-daytona@1872

@tanstack/ai-sandbox-docker

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-docker@1872

@tanstack/ai-sandbox-e2b

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-e2b@1872

@tanstack/ai-sandbox-local-process

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-local-process@1872

@tanstack/ai-sandbox-railway

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-railway@1872

@tanstack/ai-sandbox-sprites

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-sprites@1872

@tanstack/ai-sandbox-upstash-box

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-upstash-box@1872

@tanstack/ai-sandbox-vercel

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-vercel@1872

@tanstack/ai-skills

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-skills@1872

@tanstack/ai-solid

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-solid@1872

@tanstack/ai-svelte

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-svelte@1872

@tanstack/ai-typesafe

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-typesafe@1872

@tanstack/ai-utils

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-utils@1872

@tanstack/ai-vercel-gateway

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vercel-gateway@1872

@tanstack/ai-vertex

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vertex@1872

@tanstack/ai-vue

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vue@1872

@tanstack/ai-worldlabs

npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-worldlabs@1872

@tanstack/openai-base

npm i https://pkg.pr.new/TanStack/ai/@tanstack/openai-base@1872

@tanstack/preact-ai-devtools

npm i https://pkg.pr.new/TanStack/ai/@tanstack/preact-ai-devtools@1872

@tanstack/react-ai-devtools

npm i https://pkg.pr.new/TanStack/ai/@tanstack/react-ai-devtools@1872

@tanstack/solid-ai-devtools

npm i https://pkg.pr.new/TanStack/ai/@tanstack/solid-ai-devtools@1872

@tanstack/svelte-ai-devtools

npm i https://pkg.pr.new/TanStack/ai/@tanstack/svelte-ai-devtools@1872

commit: 4ffc8b1

A child resumes from its stored pending messages. The store keeps a nested child in its own thread, so those messages had no card for it, and the nested child ran again from the start. Keep the nested cards from the client's copy on the stored message of their parent tool call. Also update the harnessText test for the required threadId: one session per thread.

changeset-bot · 2026-10-22T08:18:48Z

⚠️ No Changeset found

Latest commit: 4ffc8b1

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes changesets to release 8 packages

Name	Type
@tanstack/ai-harness	Major
@tanstack/ai-acp	Patch
@tanstack/ai-code-mode	Patch
@tanstack/ai-computer-use	Patch
@tanstack/ai-harness-cli	Patch
@tanstack/ai-mcp	Patch
@tanstack/ai-sandbox	Patch
@tanstack/ai-skills	Patch

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

…twin to the chat-only conformance The stopgap in spawn.ts copied nested subagent cards from the client, which trusts the client's messages for a nested child. #1878 fixes the nested resume from the store. Until then, skip the subagent-waits test that needs it. The chat-only conformance test from main had no twin. On this branch memoryPersistence has a log store, so the log-across-instances cases need one.

AlemTuzlak approved these changes Oct 22, 2026

View reviewed changes

tombeckenham merged commit c121f26 into feat/harness-core Oct 22, 2026
13 of 14 checks passed

tombeckenham deleted the 1796-ai-harness-a-forwarded-command-completes-with-no-result-also-when-the-owner-rejects-it branch October 22, 2026 08:48

This was referenced Oct 22, 2026

ai-harness: a forwarded command completes with no result, also when the owner rejects it #1796

Closed

fix(ai, ai-persistence): resume a waiting nested subagent from the store #1879

Merged

Navigation Menu

Repository navigation

fix(ai-harness): a forwarded command ends with the owner's result or error - #1872

fix(ai-harness): a forwarded command ends with the owner's result or error#1872
tombeckenham merged 3 commits into
feat/harness-corefrom
1796-ai-harness-a-forwarded-command-completes-with-no-result-also-when-the-owner-rejects-it

tombeckenham commented Oct 22, 2026 •

edited

Loading

Uh oh!

coderabbitai Bot commented Oct 22, 2026 •

edited

Loading

Review skipped

Uh oh!

nx-cloud Bot commented Oct 22, 2026 •

edited

Loading

Uh oh!

pkg-pr-new Bot commented Oct 22, 2026 •

edited

Loading

Uh oh!

changeset-bot Bot commented Oct 22, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

tombeckenham commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🎯 Changes

✅ Checklist

🚀 Release Impact

Root cause

Possible alternatives

Testing

Linked issues

Risk / rollback

Uh oh!

coderabbitai Bot commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review skipped

Uh oh!

nx-cloud Bot commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pkg-pr-new Bot commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

changeset-bot Bot commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

⚠️ No Changeset found

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

tombeckenham commented Oct 22, 2026 •

edited

Loading

coderabbitai Bot commented Oct 22, 2026 •

edited

Loading

nx-cloud Bot commented Oct 22, 2026 •

edited

Loading

pkg-pr-new Bot commented Oct 22, 2026 •

edited

Loading

changeset-bot Bot commented Oct 22, 2026 •

edited

Loading