Skip to content
Source
Sign inSign up
TanStack/ai

feat(ai-harness): the agent tree, with typed agent actions, limits and budgets that hold, run inheritance, shared state, durable groups, and harnessAgent as a child run #1818

MergedAlemTuzlakmerged 27 commits intofeat/harness-waits fromrefs/pull/1818/headmerged by AlemTuzlak View on GitHub

AlemTuzlakcommented

Agents that code starts are now typed runs in one run tree in the log. Limits and budgets hold across the whole tree and across restarts, children follow their parent, and a group of agents can run with a concurrency limit and pick up its children after a crash. harnessAgent runs another harness as a child run of the same session. This is PR 8 of the harness redesign stack for #1755. It targets #1817 (feat/harness-waits).

🎯 Changes

  • Typed agent actions.
    • ctx.harness.agents.start(worker, input, options?), run(worker, input, options?), and group({ concurrency?, onFailure?, key? }, (g) => [g.start(...)]). The input type comes from the worker's input schema, and the result type comes from its output.
    • Plugin ctx.agents has the same type. g.run and g.runSettled are gone: use g.start, and onFailure: 'collect' gives AgentSettled entries.
    • Start options are grouped: key, limits: { timeoutMs, maxAttempts }, budget: { maxTokens, maxCost }, onModelError.
  • The run tree. RunNode (turn, agent, tool, command) lives in the log (harness.run.node). agentRuns() reads it instead of scanning every input (B4).
  • Limits and budgets.
    • maxDepth and maxCalls count per tree. maxConcurrent counts per session for top-level starts, and a start over the limit waits as queued. The counts come from the log, so they survive a restart.
    • Every usage record has runId, and a run keeps its usage sum. A run over its budget fails with code budget. A timed-out run fails with code timeout.
    • A failed start gives its budget call back (a bug before).
  • Run inheritance (D5).
    • A background run that a turn starts keeps the turn's plugins until it settles.
    • A child's signal follows its parent: cancel, timeout, and budget pass down, and a shutdown stays a shutdown. Cancelling a turn now also cancels the background runs it started.
    • Tool-mode and routed children get a real ctx.harness and share the turn's tree budget.
  • Shared state. ctx.harness.state({ key, initial }) gives a { get, update } handle. It is kept in the log, shared by key in one session, and copied by a fork.
  • Durable groups. A concurrency gate per group. With resume: true, children get keyed input ids, so a run that runs again after a crash binds its old children instead of starting new ones.
  • Routed agents take steers, late joins, beforeFinish, and onModelError, and each answer stays in the transcript.
  • harnessAgent as a child run.
    • Inside a session, it opens a child session of the other harness on <root>/agent/<run id>, in the root log.
    • It runs for the parent's user and keys, its usage counts against the parent run's budget, and its waits show at the root.
    • All its events reach the parent with parentRunId. host.open refuses a child thread id.
    • Outside a session it keeps its in-memory host.

Docs: subagents.md (agents from code, groups, limits, budgets, the run tree, shared state, routed agents, harnessAgent), plugins.md, session-log.md, usage.md, shared-logs.md, and others. The changesets are edited in place.

Decisions made while building:

  1. Keyed child ids only with resume: true. With keyed ids in every group, an onModelError retry would rebind the child that failed, so the retry could never succeed.
  2. A run that waits for a person keeps its maxConcurrent slot, the same as core does for children. A queued start does not survive a restart: the next host refuses it with expired_on_restart.
  3. onModelError works live only. A run that another host takes over logs one warning.
  4. A harnessAgent child's waits are live only. The root lists and routes them while the child session is open on this host. Another host and host.observe do not see them.

✅ 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).

Testing

Commands run:

  1. Builds of the harness and its dependencies. test:types of ai, ai-harness, ai-mcp, ai-acp, ai-harness-cli, ai-code-mode, ai-skills, and ai-sandbox. All pass.
  2. vitest related on every changed ai-harness source file, with --maxWorkers=2: 127 files, 1,785 tests, all pass after merging the latest PR 7.
  3. test:oxlint of ai-harness, pnpm test:docs, kiira check on the changed pages, and oxfmt --check on every changed file.

Not run locally: knip, E2E (including the new harness-agent-tree.spec.ts), and coverage. CI runs them.

Manual test:

  1. Run pnpm --dir packages/ai-harness exec vitest run tests/agent-group.test.ts tests/agent-limits.test.ts tests/harness-agent.test.ts.
  2. Read the host-stop test in tests/agent-group.test.ts: a group stops with its host, and the run on the next host binds its old children.
  3. Read tests/agent-limits.test.ts: a run over its token budget fails with code budget, and its parent sees the error.

How this PR makes testing easy: the new test files pin each part: run-tree, agent-limits, run-inheritance, shared-state, agent-group, and harness-agent. testing/e2e/tests/harness-agent-tree.spec.ts covers a group with a concurrency gate, a run budget, shared state, and a harnessAgent child.

Risk / rollback

High. This changes how agent runs start, count, inherit, and stop. Cancelling a turn now also cancels its background runs. The harness is not released yet, and all callers in this repo are moved. To undo, revert this PR.

Public API change

Before

const [a, b] = await ctx.agents.group(async (g) => [await g.run(research, 'x'), await g.run(write, 'y')])
await ctx.agents.start(worker, input, { timeoutMs: 60_000 })

After

const [a, b] = await ctx.harness.agents.group({ concurrency: 2 }, (g) => [g.start(research, { topic: 'x' }), g.start(write, { draft: 'y' })])
await ctx.harness.agents.start(worker, input, { limits: { timeoutMs: 60_000 }, budget: { maxTokens: 20_000 } })
const count = ctx.harness.state({ key: 'count', initial: 0 })

Review fixes

These fixes come from a stack review. #1819 to #1825 are already squash-merged into this branch. This PR is squash-merged into #1817 next.

  • harnessAgent children: the parent follows any resume of the child, including early answers and self-resumes after a wait expires. A POST run resume of a child wait streams the parent run.
  • Child ids of a resumable run use their own namespaces (/start:k:<key>, /start:n:<n>, /group:…). Keyed children never collide, and a background child that outlives its parent keeps its durable steps.
  • Budgets hold across a restart: a resumed run carries the usage of the run before it.
  • onModelError retries only the failed model call, with a backoff, at most 3 calls.
  • A plugin ctx.agents.* start inside an agent run's tool call is a child of that run automatically, so it cannot deadlock. parent: { runId } picks a run, and parent: null makes a top-level start.
  • host.open reserves only real child ids, so org/agent/notes opens. App code can never open /inner/ ids.
  • The maxConcurrent slots, group slots and tree-budget calls share one admission gate. This is a refactor with no intended behavior change.
  • harness.run.node: only the stable fields are documented, and the bookkeeping fields are @internal.

Checks: 75 ai-harness test files that cover the changed code, plus typecheck and lint. CI runs the rest.

    1. 🦋 Changeset detected

      Latest commit: c844806c8db9b745ebb348bd42dd8fa102ac40e3

      The changes in this PR will be included in the next version bump.

      This PR includes changesets to release 70 packages
      NameType
      @tanstack/ai-harness-cliMinor
      @tanstack/ai-harnessMinor
      @tanstack/ai-mcpMinor
      @tanstack/ai-persistenceMinor
      @tanstack/ai-sandboxMinor
      @tanstack/aiMinor
      @tanstack/ai-utilsMinor
      @tanstack/ai-code-modeMinor
      @tanstack/ai-acpMinor
      @tanstack/ai-skillsMinor
      @tanstack/ai-sandbox-local-processMinor
      @tanstack/ai-claude-codePatch
      @tanstack/ai-codexPatch
      @tanstack/ai-durable-streamPatch
      @tanstack/ai-grok-buildPatch
      @tanstack/ai-opencodePatch
      @tanstack/ai-sandbox-blaxelPatch
      @tanstack/ai-sandbox-boxdPatch
      @tanstack/ai-sandbox-cloudflarePatch
      @tanstack/ai-sandbox-daytonaPatch
      @tanstack/ai-sandbox-dockerPatch
      @tanstack/ai-sandbox-e2bPatch
      @tanstack/ai-sandbox-railwayPatch
      @tanstack/ai-sandbox-spritesPatch
      @tanstack/ai-sandbox-upstash-boxPatch
      @tanstack/ai-sandbox-vercelPatch
      @tanstack/ai-angularPatch
      @tanstack/ai-anthropicPatch
      @tanstack/ai-bedrockPatch
      @tanstack/ai-byteplusPatch
      @tanstack/ai-clientPatch
      @tanstack/ai-cloudflarePatch
      @tanstack/ai-code-mode-snippetsPatch
      @tanstack/ai-coherePatch
      @tanstack/ai-compactionPatch
      @tanstack/ai-devtools-corePatch
      @tanstack/ai-elevenlabsPatch
      @tanstack/ai-falPatch
      @tanstack/ai-geminiPatch
      @tanstack/ai-grokPatch
      @tanstack/ai-groqPatch
      @tanstack/ai-isolate-e2bPatch
      @tanstack/ai-llmgatewayPatch
      @tanstack/ai-lovablePatch
      @tanstack/ai-memoryPatch
      @tanstack/ai-mistralPatch
      @tanstack/ai-octanePatch
      @tanstack/ai-ollamaPatch
      @tanstack/ai-ollayaPatch
      @tanstack/ai-openaiPatch
      @tanstack/ai-openrouterPatch
      @tanstack/ai-perplexityPatch
      @tanstack/ai-preactPatch
      @tanstack/ai-reactPatch
      @tanstack/ai-reactorPatch
      @tanstack/ai-remixPatch
      @tanstack/ai-solidPatch
      @tanstack/ai-sveltePatch
      @tanstack/ai-typesafePatch
      @tanstack/ai-vercel-gatewayPatch
      @tanstack/ai-vertexPatch
      @tanstack/ai-vuePatch
      @tanstack/ai-worldlabsPatch
      @tanstack/openai-basePatch
      @tanstack/ai-isolate-cloudflarePatch
      @tanstack/ai-isolate-daytonaPatch
      @tanstack/ai-isolate-nodePatch
      @tanstack/ai-isolate-quickjs-bunPatch
      @tanstack/ai-isolate-quickjsPatch
      ag-uiPatch

      Not sure what this means? Click here to learn what changesets are.

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

    2. 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
      • Configuration used: Repository: TanStack/ai/.coderabbit.yaml
      • Review profile: CHILL
      • Plan: Advanced
      • Run ID: 0d438249-d6b2-434f-97bd-71d4fa011ab0

      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.
    3. View your CI Pipeline Execution ↗ for commit c844806c8db9b745ebb348bd42dd8fa102ac40e3

      CommandStatusDurationResult
      nx build @tanstack/ai-isolate-quickjs-bun✅ Succeeded12sView ↗

      ☁️ Nx Cloud last updated this comment at 2026-10-19 17:22:12 UTC

    4. Open in StackBlitz

      @tanstack/ai

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai@1818

      @tanstack/ai-acp

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-acp@1818

      @tanstack/ai-angular

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-angular@1818

      @tanstack/ai-anthropic

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-anthropic@1818

      @tanstack/ai-bedrock

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-bedrock@1818

      @tanstack/ai-byteplus

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-byteplus@1818

      @tanstack/ai-claude-code

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-claude-code@1818

      @tanstack/ai-client

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-client@1818

      @tanstack/ai-cloudflare

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-cloudflare@1818

      @tanstack/ai-code-mode

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-code-mode@1818

      @tanstack/ai-code-mode-snippets

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-code-mode-snippets@1818

      @tanstack/ai-codex

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-codex@1818

      @tanstack/ai-cohere

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-cohere@1818

      @tanstack/ai-compaction

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-compaction@1818

      @tanstack/ai-devtools-core

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-devtools-core@1818

      @tanstack/ai-durable-stream

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-durable-stream@1818

      @tanstack/ai-elevenlabs

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-elevenlabs@1818

      @tanstack/ai-event-client

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-event-client@1818

      @tanstack/ai-fal

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-fal@1818

      @tanstack/ai-gemini

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-gemini@1818

      @tanstack/ai-grok

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-grok@1818

      @tanstack/ai-grok-build

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-grok-build@1818

      @tanstack/ai-groq

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-groq@1818

      @tanstack/ai-harness

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-harness@1818

      @tanstack/ai-harness-cli

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-harness-cli@1818

      @tanstack/ai-isolate-cloudflare

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-cloudflare@1818

      @tanstack/ai-isolate-daytona

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-daytona@1818

      @tanstack/ai-isolate-e2b

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-e2b@1818

      @tanstack/ai-isolate-node

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-node@1818

      @tanstack/ai-isolate-quickjs

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-quickjs@1818

      @tanstack/ai-isolate-quickjs-bun

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-isolate-quickjs-bun@1818

      @tanstack/ai-llmgateway

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-llmgateway@1818

      @tanstack/ai-lovable

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-lovable@1818

      @tanstack/ai-mcp

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-mcp@1818

      @tanstack/ai-memory

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-memory@1818

      @tanstack/ai-mistral

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-mistral@1818

      @tanstack/ai-octane

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-octane@1818

      @tanstack/ai-ollama

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-ollama@1818

      @tanstack/ai-ollaya

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-ollaya@1818

      @tanstack/ai-openai

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-openai@1818

      @tanstack/ai-opencode

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-opencode@1818

      @tanstack/ai-openrouter

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-openrouter@1818

      @tanstack/ai-perplexity

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-perplexity@1818

      @tanstack/ai-persistence

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-persistence@1818

      @tanstack/ai-preact

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-preact@1818

      @tanstack/ai-react

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-react@1818

      @tanstack/ai-react-ui

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-react-ui@1818

      @tanstack/ai-reactor

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-reactor@1818

      @tanstack/ai-remix

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-remix@1818

      @tanstack/ai-sandbox

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox@1818

      @tanstack/ai-sandbox-blaxel

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-blaxel@1818

      @tanstack/ai-sandbox-boxd

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-boxd@1818

      @tanstack/ai-sandbox-cloudflare

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-cloudflare@1818

      @tanstack/ai-sandbox-daytona

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-daytona@1818

      @tanstack/ai-sandbox-docker

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-docker@1818

      @tanstack/ai-sandbox-e2b

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-e2b@1818

      @tanstack/ai-sandbox-local-process

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-local-process@1818

      @tanstack/ai-sandbox-railway

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-railway@1818

      @tanstack/ai-sandbox-sprites

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-sprites@1818

      @tanstack/ai-sandbox-upstash-box

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-upstash-box@1818

      @tanstack/ai-sandbox-vercel

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-sandbox-vercel@1818

      @tanstack/ai-skills

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-skills@1818

      @tanstack/ai-solid

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-solid@1818

      @tanstack/ai-solid-ui

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-solid-ui@1818

      @tanstack/ai-svelte

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-svelte@1818

      @tanstack/ai-typesafe

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-typesafe@1818

      @tanstack/ai-utils

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-utils@1818

      @tanstack/ai-vercel-gateway

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vercel-gateway@1818

      @tanstack/ai-vertex

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vertex@1818

      @tanstack/ai-vue

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vue@1818

      @tanstack/ai-vue-ui

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-vue-ui@1818

      @tanstack/ai-worldlabs

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/ai-worldlabs@1818

      @tanstack/openai-base

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/openai-base@1818

      @tanstack/preact-ai-devtools

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/preact-ai-devtools@1818

      @tanstack/react-ai-devtools

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/react-ai-devtools@1818

      @tanstack/solid-ai-devtools

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/solid-ai-devtools@1818

      @tanstack/svelte-ai-devtools

      npm i https://pkg.pr.new/TanStack/ai/@tanstack/svelte-ai-devtools@1818

      commit: c844806

    5. Coverage

      ✅ Coverage held across 70 compared package(s).

      Each package is measured twice in this job — on this PR and on its merge-base with main — and compared. A metric fails only when it is at or above 60% on the merge-base and under 60% on this PR. Other drops are listed but do not fail. Packages your PR doesn't affect are not measured and not listed.

      packagestatementsbranchesfunctionslines
      ai91.66% (+0.02)85.33% (-0.01)92.43% (+0.01)93.30% (+0.02)ok
      ai-acp82.89% (+0.29)71.08% (-0.02)78.82% (+0.58)84.91% (+0.37)ok
      ai-angular74.10% (0.00)62.20% (0.00)61.65% (0.00)75.60% (0.00)ok
      ai-anthropic92.00% (0.00)86.73% (0.00)95.57% (0.00)92.41% (0.00)ok
      ai-bedrock90.11% (0.00)82.82% (0.00)94.20% (0.00)91.33% (0.00)ok
      ai-byteplus97.06% (0.00)91.66% (0.00)98.27% (0.00)98.49% (0.00)ok
      ai-claude-code88.93% (0.00)76.88% (0.00)86.88% (0.00)90.40% (0.00)ok
      ai-client87.96% (0.00)80.38% (0.00)87.13% (0.00)90.26% (0.00)ok
      ai-cloudflare85.55% (0.00)76.44% (0.00)85.48% (0.00)86.79% (0.00)ok
      ai-code-mode93.12% (+0.41)86.17% (+0.51)89.70% (+0.64)93.82% (+0.31)ok
      ai-code-mode-snippets87.13% (0.00)75.67% (0.00)91.11% (0.00)88.39% (0.00)ok
      ai-codex94.55% (0.00)85.10% (0.00)89.28% (0.00)95.56% (0.00)ok
      ai-cohere88.97% (0.00)88.37% (0.00)85.71% (0.00)91.80% (0.00)ok
      ai-compaction97.38% (0.00)92.79% (0.00)100.00% (0.00)98.88% (0.00)ok
      ai-devtools83.13% (0.00)72.77% (0.00)87.07% (0.00)85.22% (0.00)ok
      ai-durable-stream86.70% (0.00)82.25% (0.00)92.30% (0.00)88.72% (0.00)ok
      ai-elevenlabs74.68% (0.00)69.59% (0.00)59.13% (0.00)75.46% (0.00)ok
      ai-event-client56.83% (0.00)46.87% (0.00)68.96% (0.00)57.89% (0.00)ok
      ai-fal91.10% (0.00)82.63% (0.00)92.40% (0.00)91.60% (0.00)ok
      ai-gemini68.33% (0.00)67.88% (0.00)64.62% (0.00)68.93% (0.00)ok
      ai-grok48.85% (0.00)53.49% (0.00)56.46% (0.00)49.67% (0.00)ok
      ai-grok-build79.19% (0.00)63.56% (0.00)77.39% (0.00)81.02% (0.00)ok
      ai-groq52.54% (0.00)47.73% (0.00)67.39% (0.00)54.66% (0.00)ok
      ai-harness93.49% (-0.04)87.58% (+0.05)93.43% (-0.13)95.35% (+0.11)ok
      ai-harness-cli91.97% (+1.16)82.74% (+0.05)88.00% (+1.52)93.16% (+1.25)ok
      ai-isolate-cloudflare94.11% (0.00)70.66% (0.00)100.00% (0.00)93.96% (0.00)ok
      ai-isolate-daytona90.25% (0.00)80.59% (0.00)100.00% (0.00)90.15% (0.00)ok
      ai-isolate-e2b92.11% (0.00)81.48% (0.00)97.36% (0.00)95.05% (0.00)ok
      ai-isolate-node84.00% (0.00)43.39% (0.00)100.00% (0.00)83.78% (0.00)ok
      ai-isolate-quickjs89.07% (0.00)76.41% (0.00)90.90% (0.00)90.30% (0.00)ok
      ai-isolate-quickjs-bun7.76% (0.00)8.08% (0.00)5.00% (0.00)8.33% (0.00)ok
      ai-llmgateway97.67% (0.00)100.00% (0.00)88.88% (0.00)97.67% (0.00)ok
      ai-lovable73.21% (0.00)55.63% (0.00)85.07% (0.00)73.96% (0.00)ok
      ai-mcp96.04% (-0.62)90.90% (-0.42)96.34% (-0.75)98.16% (-0.36)ok
      ai-memory73.87% (0.00)52.70% (0.00)79.74% (0.00)76.04% (0.00)ok
      ai-mistral84.82% (0.00)79.88% (0.00)92.59% (0.00)85.46% (0.00)ok
      ai-octane90.67% (0.00)67.74% (0.00)86.66% (0.00)90.79% (0.00)ok
      ai-ollama92.73% (0.00)78.15% (0.00)90.62% (0.00)92.68% (0.00)ok
      ai-ollaya86.84% (0.00)89.28% (0.00)85.71% (0.00)91.66% (0.00)ok
      ai-openai65.34% (0.00)65.72% (0.00)66.66% (0.00)65.63% (0.00)ok
      ai-opencode69.04% (0.00)60.82% (0.00)65.26% (0.00)70.87% (0.00)ok
      ai-openrouter87.88% (0.00)73.00% (0.00)90.78% (0.00)89.79% (0.00)ok
      ai-perplexity91.34% (0.00)90.43% (0.00)100.00% (0.00)93.87% (0.00)ok
      ai-persistence91.23% (+0.04)78.00% (0.00)97.57% (+0.01)95.70% (+0.03)ok
      ai-preact85.62% (0.00)70.81% (0.00)85.96% (0.00)87.02% (0.00)ok
      ai-react72.95% (0.00)51.89% (0.00)73.77% (0.00)73.07% (0.00)ok
      ai-reactor94.91% (0.00)83.33% (0.00)93.33% (0.00)96.55% (0.00)ok
      ai-remix36.70% (0.00)22.69% (0.00)34.18% (0.00)39.04% (0.00)ok
      ai-sandbox78.44% (+0.42)83.18% (+0.08)75.46% (+0.13)79.00% (+0.43)ok
      ai-sandbox-blaxel87.64% (0.00)77.18% (0.00)82.26% (0.00)89.71% (0.00)ok
      ai-sandbox-boxd95.86% (0.00)87.12% (0.00)95.31% (0.00)98.17% (0.00)ok
      ai-sandbox-cloudflare67.91% (0.00)57.47% (0.00)71.79% (0.00)69.62% (0.00)ok
      ai-sandbox-daytona82.70% (0.00)81.89% (0.00)67.14% (0.00)85.18% (0.00)ok
      ai-sandbox-docker80.15% (0.00)71.49% (0.00)77.38% (0.00)82.13% (0.00)ok
      ai-sandbox-e2b94.42% (0.00)88.97% (0.00)93.24% (0.00)98.63% (0.00)ok
      ai-sandbox-local-process68.61% (+0.61)53.59% (0.00)77.01% (+1.15)69.58% (+0.70)ok
      ai-sandbox-railway94.75% (0.00)92.68% (0.00)90.90% (0.00)97.59% (0.00)ok
      ai-sandbox-sprites73.23% (0.00)59.70% (0.00)73.83% (0.00)76.17% (0.00)ok
      ai-sandbox-upstash-box94.82% (0.00)91.52% (0.00)91.04% (0.00)97.76% (0.00)ok
      ai-sandbox-vercel44.67% (0.00)38.18% (0.00)35.18% (0.00)45.76% (0.00)ok
      ai-skills83.94% (+0.35)67.95% (-0.63)88.00% (+0.64)86.03% (+0.40)ok
      ai-solid52.13% (0.00)38.29% (0.00)49.82% (0.00)56.87% (0.00)ok
      ai-svelte73.76% (0.00)54.59% (0.00)69.15% (0.00)78.33% (0.00)ok
      ai-typesafe78.04% (0.00)84.61% (0.00)66.66% (0.00)82.05% (0.00)ok
      ai-utils82.81% (+7.47)75.64% (+1.64)89.47% (+3.76)84.21% (+8.83)ok
      ai-vercel-gateway75.97% (0.00)55.63% (0.00)85.41% (0.00)79.38% (0.00)ok
      ai-vertex96.77% (0.00)97.22% (0.00)100.00% (0.00)96.77% (0.00)ok
      ai-vue60.07% (0.00)34.10% (0.00)59.34% (0.00)61.04% (0.00)ok
      ai-worldlabs79.62% (0.00)71.50% (0.00)91.42% (0.00)82.66% (0.00)ok
      openai-base90.31% (0.00)78.28% (0.00)97.66% (0.00)92.36% (0.00)ok
    6. github-actions[bot](GitHub profile, opens in a new tab) added labelswaiting-on: maintainer

    7. AlemTuzlak merged this1b57d19

    8. AlemTuzlak closed this

Sign in to join this conversation.