Skip to content

feat(nodes): add auto-discovered Ollama inference#99234

Merged
steipete merged 7 commits into
mainfrom
codex/node-local-inference
Jul 3, 2026
Merged

feat(nodes): add auto-discovered Ollama inference#99234
steipete merged 7 commits into
mainfrom
codex/node-local-inference

Conversation

@steipete

@steipete steipete commented Jul 2, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Closes #99228

AI-assisted: Codex. I understand and reviewed the implementation and its node/plugin boundaries.

What Problem This Solves

Agents connected to desktop or server nodes could not discover and use a fast local model on that node. Using Ollama as the Gateway provider requires direct Gateway network access, while shelling out through node exec has no typed model discovery, capability filtering, or bounded result contract.

Operators also need node-level control: Ollama may be running on a node without being available to agents.

Why This Change Was Made

The bundled Ollama plugin now owns two node-host commands (ollama.models and ollama.chat) plus a node_inference agent tool. It reuses the existing plugin command registry, node invoke policy, and runtime node APIs; Ollama remains loopback-only on the node, with no new core protocol, dependency, or remote-provider fallback.

Startup activation is deliberate: node-host command metadata must be registered before a node advertises its capabilities, so lazy Gateway provider discovery cannot expose these commands. Ollama API requests remain demand-driven.

Discovery positively verifies completion capability, excludes cloud and embedding-only rows, and ranks loaded models first. Runs validate one exact local model, bound prompts/output/timeouts/error bodies, report usage/timings, and surface token-limit truncation as an error.

plugins.entries.ollama.config.nodeInference.enabled=false disables command registration on the node host while leaving Ollama itself, the Gateway provider, and the agent tool unchanged. This gives each node host an explicit opt-out without expanding core config.

The final hosted artifact gate also exposed a deterministic tsdown OOM at its existing 8 GB V8 heap ceiling. The workflow now gives only that build step a 12 GB ceiling on its 16-vCPU runner; the exact-head rerun passed.

User Impact

Users can keep small helper tasks on a paired node: start Ollama and the OpenClaw node host, approve the advertised commands, then ask an agent to discover and run an installed model. The path is available on macOS, Linux, and Windows node hosts and stays subject to normal node pairing and command policy.

The Ollama provider guide documents setup, discovery/filtering, security/locality, direct command verification, the node-level opt-out, re-enable behavior, and foreground-vs-service restart. The Nodes guide links to that workflow.

Evidence

  • Real OpenAI-to-local E2E: OpenAI agent turn -> Gateway -> paired node host -> local Ollama qwen3:0.6b; two node_inference calls returned NODE_LOCAL_OPENAI_E2E_OK with no fallback, and local generation completed in 1.61 seconds.
  • Real node opt-out: with nodeInference.enabled=false, Ollama remained reachable with 17 models while the paired node advertised neither the local-inference capability nor ollama.models / ollama.chat.
  • Live Ollama Cloud API-key suite: 2 passed, 2 intentionally skipped.
  • Focused regression proof: 497 feature tests, 80 Gateway tests, and 103 node-host/plugin/Ollama opt-out assertions passed.
  • Final CI regression reproduction: node scripts/run-vitest.mjs src/security/audit-plugin-readonly-scope.test.ts src/security/audit-node-command-findings.test.ts (7/7 passed).
  • Build: pnpm build.
  • Docs: MDX passed for 688 files; 5,545 links checked with 0 broken; docs formatting and i18n glossary checks passed.
  • Remote changed-surface check: Blacksmith Testbox tbx_01kwkbn7fvsm770h1jm75thbtm, exit 0; delegated Actions run. Covered extension production/test types, all core and plugin lint shards, guards, and import-cycle analysis.
  • Earlier full changed-surface check: Blacksmith Testbox tbx_01kwk0vhy7st9eshebcc7y3fh7, exit 0; delegated Actions run.
  • Artifact-build CI: the 8 GB tsdown OOM reproduced twice; raising only that step to 12 GB passed on exact-head run 28647141731.
  • Exact-head CI: run 28647141655 passed all relevant Node, QA, type, lint, guard, build, docs, CodeQL, and Opengrep checks.
  • Fresh full-branch autoreview: .agents/skills/autoreview/scripts/autoreview --mode branch --base origin/main --stream-engine-output returned clean with no accepted/actionable findings.

@openclaw-barnacle openclaw-barnacle Bot added docs Improvements or additions to documentation plugin: file-transfer extensions: ollama size: L maintainer Maintainer-authored PR labels Jul 2, 2026
@clawsweeper

clawsweeper Bot commented Jul 2, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Codex review: needs real behavior proof before merge. Reviewed July 3, 2026, 4:11 AM ET / 08:11 UTC.

Summary
The PR adds Ollama node-host discovery/chat commands, a node_inference agent tool, startup activation with a node-level opt-out, docs/tests, and broad related app, i18n, dependency, workflow, channel, and UI changes.

Reproducibility: not applicable. this is a feature PR rather than a bug reproduction. Source inspection confirms current main lacks the requested node_inference tool and Ollama node-host commands.

Review metrics: 3 noteworthy metrics.

  • Config/default surfaces: 1 startup default changed, 1 plugin config option added. Ollama startup activation and nodeInference.enabled can change upgrade and operator behavior before users opt in or out.
  • Security-sensitive side surfaces: 6 dependency/workspace manifests plus 1 CodeQL workflow changed. Supply-chain and security automation changes need owner review separate from the provider feature logic.
  • Post-proof commits: 3 commits after the named proven head. The live head includes later changes after the real behavior proof comment, so proof no longer maps cleanly to the exact branch under review.

Stored data model
Persistent data-model change detected: serialized state: ui/src/ui/chat/session-controls.ts, serialized state: ui/src/ui/controllers/sessions.ts, unknown-data-model-change: ui/src/ui/chat/session-controls.ts, unknown-data-model-change: ui/src/ui/controllers/sessions.ts, unknown-truncated-pull-files. Confirm migration or upgrade compatibility proof before merge.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #99228
Summary: This PR is the candidate implementation for the open node-local Ollama inference feature request; the older local-provider request is adjacent because it focused on Gateway provider setup rather than paired-node loopback execution.

Members:

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Merge readiness
Overall: 🧂 unranked krab
Proof: 🦪 silver shellfish
Patch quality: 🧂 unranked krab
Result: blocked until stronger real behavior proof is added.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

  • Drop or split unrelated dependency, workflow, app, UI, channel, i18n, and changelog changes.
  • Update the PR body with redacted exact-current-head paired-node Ollama proof for discovery, chat, and nodeInference.enabled=false; a fresh ClawSweeper review should run after that, or a maintainer can comment @clawsweeper re-review.
  • Get explicit maintainer/security review for the startup default, config option, node command boundary, dependency graph, and CodeQL workflow changes if any remain.

Proof guidance:

  • [P1] Needs stronger real behavior proof before merge: The PR has strong copied live-output claims for an earlier head, but it does not show the paired-node Ollama and opt-out behavior after the current head 873298d57484b30033b05ce433bb334887ca9482; add redacted current-head terminal/log/recording proof in the PR body before merge. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Mantis proof suggestion
The PR includes visible Telegram progress/draft-stream changes, and native Telegram Desktop proof would help reviewers see whether draft/final delivery and progress lanes still behave correctly. A maintainer can ask Mantis to capture proof by posting this exact PR comment:

@openclaw-mantis telegram desktop proof: verify progress-mode draft/final delivery, reasoning/commentary lanes, and no duplicate or stale preview messages after this PR.

Risk before merge

  • [P1] The branch mixes the intended Ollama node-inference feature with unrelated dependency/workspace, CodeQL workflow, mobile app, UI/i18n, Telegram/channel, docs, and CHANGELOG changes, so maintainers cannot review the feature in isolation.
  • [P1] Changing Ollama activation.onStartup to true and adding plugins.entries.ollama.config.nodeInference.enabled can affect startup, node-host advertising, and upgrade/operator behavior for existing users.
  • [P1] The dependency graph, shrinkwraps, pnpm-workspace.yaml, and a CodeQL workflow change are supply-chain and automation surfaces that need separate owner/security review or removal before merge.
  • [P1] The real paired-node Ollama proof is tied to an earlier head SHA, while the live head has later commits; exact-current-head proof is still needed before a pass verdict.

Maintainer options:

  1. Narrow and re-prove the branch (recommended)
    Drop or split unrelated app, dependency, workflow, generated-output, channel/UI, and changelog changes, then rerun exact-current-head paired-node Ollama proof and focused CI on the reduced diff.
  2. Accept broad scope with owner review
    If maintainers intentionally want the expanded branch, route dependency, CodeQL workflow, mobile app, UI, and channel/plugin changes through their owners and accept the upgrade/security impact explicitly.
  3. Pause this landing candidate
    If the branch cannot be reduced to a reviewable Ollama node-inference diff, pause or close it and preserve the useful implementation in a new focused PR.

Next step before merge

  • [P1] The protected maintainer label plus unresolved branch scope, supply-chain, workflow, security-boundary, proof, and compatibility decisions require human maintainer handling rather than an automated repair lane.

Security
Needs attention: The intended Ollama node code did not show a concrete secret leak in the inspected paths, but the live head includes unscoped supply-chain and security-workflow changes that need removal or dedicated owner review.

Review findings

  • [P1] Keep this branch scoped to Ollama node inference — pnpm-workspace.yaml:10
Review details

Best possible solution:

Restore a narrow Ollama node-inference PR, or explicitly route every non-Ollama surface through the relevant owners, then provide exact-current-head paired-node Ollama proof plus fresh-install and upgrade evidence for the startup/config behavior.

Do we have a high-confidence way to reproduce the issue?

Not applicable; this is a feature PR rather than a bug reproduction. Source inspection confirms current main lacks the requested node_inference tool and Ollama node-host commands.

Is this the best way to solve the issue?

No for the current live head. The Ollama-owned node-command/tool shape is a plausible owner-boundary fit, but the branch is too broad and needs exact-current-head proof plus maintainer acceptance of the startup/config behavior before it is the best landing path.

Full review comments:

  • [P1] Keep this branch scoped to Ollama node inference — pnpm-workspace.yaml:10
    The live head mixes the Ollama feature with unrelated dependency/workspace changes, app/i18n/UI/channel churn, a CodeQL workflow edit, and release-owned changelog content. Please drop or split that churn before merge so maintainers are not reviewing supply-chain and app behavior under a provider feature PR.
    Confidence: 0.93

Overall correctness: patch is incorrect
Overall confidence: 0.91

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 5e61da3deb19.

Label changes

Label changes:

  • add mantis: telegram-visible-proof: Mantis should capture Telegram visible proof. The PR changes visible Telegram progress/draft-stream behavior in production Telegram files, so a short Telegram Desktop proof would materially help if those changes remain in this branch.

Label justifications:

  • P2: The intended feature is a bounded but meaningful local Ollama/node workflow improvement, while the current blocker is merge readiness rather than an active production regression.
  • merge-risk: 🚨 compatibility: The PR changes Ollama startup activation and adds a plugin config/default surface that can affect existing node-host startup and upgrade behavior.
  • merge-risk: 🚨 security-boundary: The feature exposes node-local model execution through paired node commands and the live diff also changes dependency and lockfile surfaces.
  • merge-risk: 🚨 automation: The live diff changes a CodeQL critical-quality workflow and CI artifact heap behavior, so automation behavior cannot be assumed from ordinary feature tests.
  • rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🦪 silver shellfish and patch quality is 🧂 unranked krab.
  • status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs stronger real behavior proof before merge: The PR has strong copied live-output claims for an earlier head, but it does not show the paired-node Ollama and opt-out behavior after the current head 873298d57484b30033b05ce433bb334887ca9482; add redacted current-head terminal/log/recording proof in the PR body before merge. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
  • mantis: telegram-visible-proof: Mantis should capture Telegram visible proof. The PR changes visible Telegram progress/draft-stream behavior in production Telegram files, so a short Telegram Desktop proof would materially help if those changes remain in this branch.
Evidence reviewed

Security concerns:

  • [medium] Unscoped dependency graph changes — pnpm-workspace.yaml:10
    The PR changes pnpm-workspace.yaml, lock/shrinkwrap files, and package manifests that are unrelated to the stated Ollama node-inference feature and are supply-chain review surfaces.
    Confidence: 0.9
  • [low] Security workflow change mixed into feature branch — .github/workflows/codeql-critical-quality.yml:449
    The PR changes the CodeQL critical-quality workflow filter even though the PR purpose is Ollama node inference, so security automation behavior should not change here without owner review.
    Confidence: 0.84

What I checked:

  • Repository policy read and applied: Read the root AGENTS.md plus scoped guides for extensions, plugin SDK, plugins, gateway, agents, UI, Android, docs, and scripts; the applicable guidance treats plugin/config/default, dependency, workflow, and security-boundary changes as compatibility-sensitive review surfaces. (AGENTS.md:1, 5e61da3deb19)
  • Current main lacks the requested node inference surface: Current main has Ollama startup activation disabled, no node_inference tool contract, and no ollama.models, ollama.chat, local-inference, or nodeInference matches in the searched Ollama, agent, Gateway, node-host, and docs paths. (extensions/ollama/openclaw.plugin.json:5, 5e61da3deb19)
  • Existing plugin/node seams already exist on main: Current main already exposes plugin-owned node-host commands and node invoke policies through the plugin API, plus deterministic node-host command listing and dispatch; the PR builds on this existing boundary rather than adding a new core protocol. (src/plugins/types.ts:2205, 5e61da3deb19)
  • PR changes Ollama startup/config defaults: The PR changes Ollama activation.onStartup from false to true, adds contracts.tools: ["node_inference"], and adds a nodeInference.enabled plugin config surface and UI hints. (extensions/ollama/openclaw.plugin.json:5, 873298d57484)
  • PR implementation adds node-local inference tool path: The new Ollama module registers ollama.models, ollama.chat, a desktop/server pass-through node invoke policy, and a node_inference agent tool that calls api.runtime.nodes.invoke with operator.write scope. (extensions/ollama/src/node-inference.ts:371, 873298d57484)
  • Broad branch scope and supply-chain surface: The live PR changes pnpm-workspace.yaml, multiple lock/shrinkwrap/package files, a CodeQL workflow, mobile app/license/i18n files, Telegram/channel code, UI files, docs, and CHANGELOG content alongside the Ollama feature. (pnpm-workspace.yaml:10, 873298d57484)

Likely related people:

  • steipete: Recent GitHub path history shows repeated Ollama/provider and node/Gateway policy work, including Ollama cloud discovery and node policy docs, and this linked issue/PR are also authored by this contributor. (role: recent area contributor and feature requester; confidence: high; commits: 11484f8a1483, af79cd6a9d35, 2da49ef4acbf; files: extensions/ollama/index.ts, extensions/ollama/src/provider-models.ts, src/gateway/node-command-policy.ts)
  • Alix-007: Recently hardened bounded Ollama model-discovery response reads in the same provider-model discovery path reused by the new node model discovery implementation. (role: recent Ollama discovery contributor; confidence: medium; commits: d1c2934d0d17; files: extensions/ollama/src/provider-models.ts)
  • obviyus: Recent history includes Android node command work and plugin-SDK command behavior, both adjacent to node-host command advertising and invocation policy. (role: adjacent node and plugin-SDK contributor; confidence: medium; commits: 3d1ec371295c, a5e11c0083c5, e99b323a6b40; files: src/gateway/node-command-policy.ts, src/plugins/types.ts, apps/android/app/src/main/java/ai/openclaw/app/NodeRuntime.kt)
  • vincentkoc: Recent history touches plugin registry/type surfaces and provider catalog routing near the runtime boundary this PR extends. (role: recent plugin/provider runtime contributor; confidence: medium; commits: 74e7b8d47b18, d07ba5f2656d, 27b15a19e84c; files: src/plugins/types.ts, extensions/ollama/index.ts, src/plugins/runtime.ts)
What the crustacean ranks mean
  • 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
  • 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
  • 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
  • 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
  • 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
  • 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
  • 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works
  • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
  • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
  • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
  • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
  • Maintainers can also comment @clawsweeper review to request a fresh review only.
  • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
  • Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
  • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

@clawsweeper clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. feature: ✨ showcase ClawSweeper spotlight: unusually compelling feature idea for maintainer attention. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. P2 Normal backlog priority with limited blast radius. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. labels Jul 2, 2026
@openclaw-barnacle openclaw-barnacle Bot added gateway Gateway runtime agents Agent runtime and tooling size: XL and removed size: L labels Jul 3, 2026
@github-actions github-actions Bot mentioned this pull request Jul 3, 2026
Open
@steipete steipete self-assigned this Jul 3, 2026
@steipete
steipete requested a review from a team as a code owner July 3, 2026 07:18
@steipete

steipete commented Jul 3, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

Land-ready maintainer verification

Exact head: 873298d57484b30033b05ce433bb334887ca9482.

  • Maintainer decision: accept default Ollama startup activation and node-local model execution, with the documented per-node nodeInference.enabled=false opt-out.
  • Before/after: current main has no typed node-owned Ollama inference path. The PR's real OpenAI agent -> Gateway -> paired node -> local Ollama run made two tool calls, returned NODE_LOCAL_OPENAI_E2E_OK, and used no fallback. With opt-out enabled, Ollama remained reachable with 17 models while the node advertised neither the local-inference capability nor Ollama commands.
  • Focused proof: 497 feature tests, 80 Gateway tests, 103 opt-out assertions; node scripts/run-vitest.mjs src/security/audit-plugin-readonly-scope.test.ts src/security/audit-node-command-findings.test.ts passed 7/7; pnpm build passed.
  • Live proof: Ollama Cloud API-key suite passed 2/2 runnable cases with 2 intended skips; real local qwen3:0.6b generation completed in 1.61 seconds.
  • Remote proof: Testbox tbx_01kwkbn7fvsm770h1jm75thbtm (run 28643583515) and tbx_01kwk0vhy7st9eshebcc7y3fh7 (run 28636739318), both exit 0.
  • Artifact-build gate: the existing 8 GB tsdown heap ceiling OOMed twice; raising only that step to 12 GB passed run 28647141731.
  • Exact-head CI: run 28647141655 is terminal green, including QA, Node, types, lint, guards, build, docs, CodeQL, and Opengrep.
  • Docs: provider setup, discovery/ranking, locality/security, direct verification, node-level disable/re-enable, and restart behavior are documented; MDX, 5,545-link, format, glossary, and docs-map checks passed.
  • Fresh full-branch autoreview: clean; no accepted/actionable findings. Native review artifacts validated with zero findings; prepare guards passed.

Known proof gaps: none.

@openclaw-barnacle openclaw-barnacle Bot added channel: discord Channel integration: discord channel: imessage Channel integration: imessage channel: line Channel integration: line channel: matrix Channel integration: matrix channel: msteams Channel integration: msteams labels Jul 3, 2026
@github-actions github-actions Bot added the dependencies-changed PR changes dependency-related files label Jul 3, 2026
@github-actions

github-actions Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

Dependency Guard

This PR changes dependency-related files. Maintainers should confirm these changes are intentional.

Changed files:

  • extensions/memory-lancedb/npm-shrinkwrap.json
  • extensions/msteams/npm-shrinkwrap.json
  • extensions/qa-lab/package.json
  • extensions/slack/npm-shrinkwrap.json
  • pnpm-lock.yaml
  • pnpm-workspace.yaml

Maintainer follow-up:

  • Review whether the dependency changes are intentional.
  • Inspect resolved package deltas when lockfile, shrinkwrap, or workspace dependency policy changes are present.
  • Treat package-lock.json and npm-shrinkwrap.json diffs as security-review surfaces.
  • Run pnpm deps:changes:report -- --base-ref origin/main --markdown /tmp/dependency-changes.md --json /tmp/dependency-changes.json locally for detailed release-style evidence.

@github-actions

github-actions Bot commented Jul 3, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Dependency graph changes noted

This PR includes dependency graph changes. The dependency guard is informational because the PR author is a repository admin or a member of @openclaw/openclaw-secops.

  • Current SHA: 873298d57484b30033b05ce433bb334887ca9482
  • Trusted actor: @steipete
  • Trusted role: pull request author; openclaw-secops

Security review is still recommended before merge when the dependency graph change is intentional.

@clawsweeper clawsweeper Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label Jul 3, 2026
@steipete

steipete commented Jul 3, 2026

Copy link
Copy Markdown
Contributor Author

Merged via squash.

@NianJiuZst NianJiuZst mentioned this pull request Jul 3, 2026
Closed
@openclaw-mantis

openclaw-mantis Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

Mantis Telegram Desktop Proof

Summary: Mantis captured native Telegram Desktop before/after GIF evidence for a tool-using chat turn.

Main screenshot This PR screenshot
Baseline native Telegram Desktop screenshot Candidate native Telegram Desktop screenshot
Main This PR
Baseline native Telegram Desktop proof GIF Candidate native Telegram Desktop proof GIF

Motion-trimmed clips:

Raw QA files: https://artifacts.openclaw.ai/mantis/telegram-desktop/pr-99234/run-28647614569-1/index.json

@github-actions github-actions Bot mentioned this pull request Jul 3, 2026
Open
@clawsweeper clawsweeper Bot mentioned this pull request Jul 3, 2026
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

@steipete steipete

Labels

agents Agent runtime and tooling app: android App: android app: ios App: ios app: web-ui App: web-ui channel: discord Channel integration: discord channel: feishu Channel integration: feishu channel: imessage Channel integration: imessage channel: irc channel: line Channel integration: line channel: matrix Channel integration: matrix channel: msteams Channel integration: msteams channel: qa-channel Channel integration: qa-channel channel: qqbot channel: signal Channel integration: signal channel: slack Channel integration: slack channel: telegram Channel integration: telegram channel: voice-call Channel integration: voice-call channel: whatsapp-web Channel integration: whatsapp-web channel: zalo Channel integration: zalo channel: zalouser Channel integration: zalouser cli CLI command changes commands Command implementations dependencies-changed PR changes dependency-related files docs Improvements or additions to documentation extensions: acpx extensions: anthropic-vertex extensions: brave extensions: codex extensions: codex-supervisor Extension: codex-supervisor extensions: copilot extensions: device-pair extensions: diagnostics-otel Extension: diagnostics-otel extensions: duckduckgo extensions: firecrawl extensions: google extensions: memory-core Extension: memory-core extensions: memory-lancedb Extension: memory-lancedb extensions: memory-wiki extensions: minimax extensions: moonshot extensions: nvidia extensions: ollama extensions: openai extensions: openrouter extensions: parallel extensions: perplexity extensions: policy extensions: qa-lab extensions: tavily extensions: tokenjuice Changes to the bundled tokenjuice extension extensions: xai feature: ✨ showcase ClawSweeper spotlight: unusually compelling feature idea for maintainer attention. gateway Gateway runtime maintainer Maintainer-authored PR mantis: telegram-visible-proof Mantis should capture Telegram visible proof. merge-risk: 🚨 automation 🚨 May affect CI, automerge, proof capture, label sync, or maintainer automation. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. merge-risk: 🚨 security-boundary 🚨 May affect sandboxing, authorization, credentials, or sensitive data. P2 Normal backlog priority with limited blast radius. plugin: bonjour Plugin integration: bonjour plugin: file-transfer plugin: google-meet plugin: workboard rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. scripts Repository scripts size: XL status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[Feature]: Add auto-discovered Ollama inference on nodes

1 participant

@steipete