fix(models): mark local Ollama rows available by qingminglong · Pull Request #97491 · openclaw/openclaw

qingminglong · 2026-06-28T13:06:22Z

What Problem This Solves

Fixes an issue where users running openclaw models list --json could see a reachable local Ollama model reported with local: true but available: false when registry model-level availability did not include that exact model key.

Why This Change Was Made

Local model rows already classify loopback/local baseUrl values, and the models CLI documents availability/auth as provider-level metadata from local-provider markers rather than a live per-model execution probe. This change treats local provider rows as available from that local marker while preserving registry model-key availability and provider-auth fallback for non-local rows.

User Impact

Users with local Ollama models now get model-list output that matches the local-provider auth semantics: configured local Ollama rows no longer look unavailable solely because registry discovery omitted the exact model key.

Evidence

Claim proved: a reachable local Ollama model is listed with local: true and available: true on the patched branch.
Commands / artifacts:
- docker run -d --rm --name openclaw-ollama-proof -p 11434:11434 alpine/ollama:latest -> started a real Ollama daemon reachable at http://127.0.0.1:11434.
- docker exec openclaw-ollama-proof ollama pull smollm2:135m -> pulled a local model into that daemon.
- curl.exe -sS http://127.0.0.1:11434/api/tags -> returned smollm2:135m from the daemon.
- node --import tsx src/entry.ts models list --json --provider ollama at commit b5390f5c15, with isolated OPENCLAW_HOME, isolated OPENCLAW_CONFIG_PATH, and OLLAMA_API_KEY=ollama-local -> passed.
- node scripts/run-vitest.mjs src/commands/models/list.model-row.test.ts --reporter=verbose -> passed, 4 tests.
- node scripts/run-vitest.mjs src/commands/models/list.rows.test.ts --reporter=verbose -> passed, 4 tests.
- git diff --check -> passed.
- python .agents\skills\autoreview\scripts\autoreview --mode local -> clean, no accepted/actionable findings.
Observed result: the new regression test fails before the fix with expected false to be true, then passes after local base URL rows are included in row availability.
Real Ollama CLI proof output:

COMMAND: node --import tsx src/entry.ts models list --json --provider ollama
COMMIT: b5390f5c15
OLLAMA_DAEMON: Docker Desktop container openclaw-ollama-proof, alpine/ollama:latest, reachable at http://127.0.0.1:11434
OPENCLAW_HOME: isolated temp dir
OPENCLAW_CONFIG_PATH: isolated temp openclaw.json
OLLAMA_API_KEY: ollama-local
OLLAMA_TAGS_BEFORE: {"models":[{"name":"smollm2:135m","model":"smollm2:135m","modified_at":"2026-06-28T13:57:38.596605328Z","size":270898672,"digest":"9077fe9d2ae1a4a41a868836b56b8163731a8fe16621397028c2c76f838c6907","details":{"format":"gguf","family":"llama","families":["llama"],"parameter_size":"134.52M","quantization_level":"F16","context_length":8192,"embedding_length":576},"capabilities":["completion"]}]}
EXIT: 0

{
  "count": 1,
  "models": [
    {
      "key": "ollama/smollm2:135m",
      "name": "smollm2:135m",
      "input": "text",
      "contextWindow": 8192,
      "local": true,
      "available": true,
      "tags": [],
      "missing": false
    }
  ]
}

Supporting loopback proof: before the Dockerized daemon was available, the same OpenClaw CLI entrypoint was also run against a local Ollama-compatible endpoint and returned available: true for ollama/qwen3.6:35b-a3b.
Not tested / proof gaps: I did not install native Windows Ollama or run against the user's normal local model store. The live proof above uses a real Ollama daemon running in Docker Desktop with a pulled local model and exercises OpenClaw's real CLI entrypoint, config loading, provider discovery HTTP calls, and JSON output.

clawsweeper · 2026-06-28T13:07:33Z

Codex review: needs maintainer review before merge. Reviewed June 28, 2026, 1:50 PM ET / 17:50 UTC.

Summary
The branch changes model-list row availability so local provider rows are available from the local base URL marker, adds a regression test, and currently also carries main-equivalent marketplace, provider catalog, SDK, and response-limit refresh changes from its second commit.

PR surface: Source +1376, Tests +2117, Docs +17, Generated 0, Other 0. Total +3510 across 50 files.

Reproducibility: yes. at source level: current main can render a loopback Ollama row as local while leaving available false when availableKeys omits that key. The PR body also supplies after-fix live CLI output against a Dockerized Ollama daemon.

Review metrics: 2 noteworthy metrics.

Unique Config Surface: 0 added, 0 changed, 0 removed relative to current main. The marketplace config/default files shown in the stale PR diff match current main and are not part of the unique Ollama fix.
Public Availability Rule: 1 boolean row-availability rule changed. The user-visible CLI JSON/table output for local provider rows changes, so maintainers should review the compatibility semantics before merge.

Stored data model
Persistent data-model change detected: database schema: src/cli/plugins-cli.marketplace-refresh.test.ts, database schema: src/plugins/official-external-plugin-catalog.test.ts, serialized state: extensions/minimax/video-generation-provider.test.ts, serialized state: src/config/zod-schema.marketplaces.test.ts, serialized state: src/plugin-sdk/provider-catalog-live-runtime.test.ts, serialized state: src/plugins/official-external-plugin-catalog.test.ts, and 7 more. Confirm migration or upgrade compatibility proof before merge.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #92224
Summary: This PR is the active candidate fix for the open local Ollama model-list availability bug; two earlier closed PRs attempted the same root cause but did not land.

Members:

canonical: Bug: CLI shows available: false for local Ollama models that are running and working #92224 - The open issue contains the reported local Ollama available: false behavior and is closed by this PR body.
superseded: fix(cli): mark local models (Ollama) available in models list without auth #92228 - This closed unmerged PR attempted the same local-Ollama availability fix with narrower proof.
superseded: #92224: fix(models): report local models as available in CLI list output #92363 - This closed unmerged PR attempted the same local-Ollama availability fix and is no longer the active candidate.

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

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

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

Rank-up moves:

[P2] Rebase onto current main so the final diff clearly shows only the intended model-list fix before merge.

Risk before merge

[P1] Merging intentionally changes public models list --json and table Auth/availability semantics for local-baseUrl rows from exact registry-key availability to local-marker availability; maintainers should accept that this is provider-level metadata, not live per-model readiness.
[P1] The head is behind current main and the displayed PR diff includes main-equivalent marketplace/provider/SDK/network files; rebase or merge-result review should confirm the final surface collapses to the intended model-list fix.

Maintainer options:

Rebase and land the narrow model fix (recommended)
Rebase onto current main so the review surface shows the two model-list files, then merge after required CI and maintainer acceptance of the output semantics.
Accept provider-level availability semantics
Maintainers can intentionally accept that local rows use the local marker for availability even when the endpoint is not being live-probed per model.
Pause for a live-probe design instead
If available must mean exact live Ollama readiness, pause this PR and design a separate opt-in live probe rather than merging this display fix.

Next step before merge

[P2] The remaining action is maintainer acceptance of compatibility-sensitive CLI availability semantics plus stale-head refresh, not an automated code repair.

Security
Cleared: No concrete security or supply-chain issue was found in the unique model-list fix; the broader security-sensitive files visible in the stale diff match current-main refresh content.

Review details

Best possible solution:

Rebase onto current main, keep the focused model-row fix and regression/live Ollama proof, and have a maintainer explicitly accept the local-marker availability semantics before merge.

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

Yes at source level: current main can render a loopback Ollama row as local while leaving available false when availableKeys omits that key. The PR body also supplies after-fix live CLI output against a Dockerized Ollama daemon.

Is this the best way to solve the issue?

Yes, this is the narrowest maintainable fix if available follows the documented provider-level local-marker semantics. A live per-model reachability probe would be a different feature and should not be smuggled into this bug fix.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 2e5e5e5af905.

Label changes

Label justifications:

P2: This is a normal-priority CLI display bug for local Ollama users; active sessions still work, but model-list availability is misleading.
merge-risk: 🚨 compatibility: The PR changes public model-list availability output for existing local-provider configurations and could affect users or scripts interpreting available as exact live readiness.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes after-fix live output from a real Dockerized Ollama daemon with a pulled local model and openclaw models list --json --provider ollama showing available: true.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes after-fix live output from a real Dockerized Ollama daemon with a pulled local model and openclaw models list --json --provider ollama showing available: true.

Evidence reviewed

PR surface:

Source +1376, Tests +2117, Docs +17, Generated 0, Other 0. Total +3510 across 50 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	28	1487	111	+1376
Tests	17	2165	48	+2117
Docs	3	26	9	+17
Config	0	0	0	0
Generated	1	3	3	0
Other	1	4	4	0
Total	50	3685	175	+3510

What I checked:

Repository policy read: Root and scoped AGENTS.md files for docs, extensions, plugin SDK, agents, plugins, and scripts were read; the review applied the repo guidance that config/default surfaces, plugin SDK changes, provider routing, and public CLI behavior are compatibility-sensitive. (AGENTS.md:1, 2e5e5e5af905)
Current main still has the false-negative path: At current main, modelIsAvailable only checks availableKeys, so a local row with a missing registry key can still produce available: false when fallback is disabled. (src/commands/models/list.model-row.ts:53, 2e5e5e5af905)
PR fixes the implicated row decision: The PR makes modelIsAvailable true for local base URLs before falling back to registry/provider-auth availability, matching the reported Ollama local-row case. (src/commands/models/list.model-row.ts:53, 4dd1874b04f1)
Focused regression coverage: The added test covers an Ollama loopback row whose availableKeys set omits the exact model key and expects available: true. (src/commands/models/list.model-row.test.ts:60, 4dd1874b04f1)
Documented CLI semantics: The models CLI docs define Auth as provider-level read-only metadata from local markers and other auth/config markers, not an exact per-model execution probe. Public docs: docs/cli/models.md. (docs/cli/models.md:55, 2e5e5e5af905)
Ollama local-marker contract: The Ollama docs say local and LAN hosts use the ollama-local marker instead of a real bearer token, supporting the local-marker availability interpretation. Public docs: docs/providers/ollama.md. (docs/providers/ollama.md:27, 2e5e5e5af905)

Likely related people:

shakkernerd: Recent history shows model-list row and auth-index work in the central toModelRow and row-source path. (role: feature-history contributor; confidence: high; commits: b418c08a2297, 35171f4e4715, c7af4dcb312b; files: src/commands/models/list.model-row.ts, src/commands/models/list.rows.ts)
steipete: History shows adjacent model-list docs, auth catalog, and Ollama/provider-auth semantics work around the affected behavior. (role: adjacent provider/auth contributor; confidence: medium; commits: 73752f07f219, 9d7f83b17504, d5c094f1691a; files: docs/cli/models.md, docs/providers/ollama.md, src/commands/models/list.rows.ts)
vincentkoc: The PR is assigned to this user, and recent history shows broad current-main touches in model-list row/source files plus the current branch refresh commit. (role: recent area contributor and assigned reviewer; confidence: medium; commits: eea777c9fc9c, 4dd1874b04f1; files: src/commands/models/list.rows.ts, src/commands/models/list.model-row.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.

qingminglong · 2026-06-28T13:36:37Z

@clawsweeper re-review

I updated the PR body with additional CLI evidence from commit b5390f5c15:

node --import tsx src/entry.ts models list --json --provider ollama
isolated OPENCLAW_HOME and OPENCLAW_CONFIG_PATH
loopback Ollama-compatible endpoint hit by real OpenClaw provider discovery: GET /api/tags, POST /api/show
output row: ollama/qwen3.6:35b-a3b, local: true, available: true

The body also keeps the remaining proof gap explicit: this Windows host does not currently have a real Ollama daemon, ollama CLI, or running Docker daemon, so this is loopback-compatible endpoint proof rather than full real-daemon Ollama proof.

clawsweeper · 2026-06-28T13:36:40Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

qingminglong · 2026-06-28T14:01:17Z

@clawsweeper re-review

I updated the PR body with real Ollama daemon proof for commit b5390f5c15.

New evidence added:

real Ollama daemon running in Docker Desktop: openclaw-ollama-proof, alpine/ollama:latest, reachable at http://127.0.0.1:11434
pulled local model: smollm2:135m
curl.exe -sS http://127.0.0.1:11434/api/tags returned that model from the daemon
patched branch command: node --import tsx src/entry.ts models list --json --provider ollama
isolated OPENCLAW_HOME, isolated OPENCLAW_CONFIG_PATH, and OLLAMA_API_KEY=ollama-local
output row: ollama/smollm2:135m, local: true, available: true, EXIT: 0

Remaining gap is only that I did not install native Windows Ollama or use the user's normal model store; this is a real Dockerized Ollama daemon with a pulled local model, not the earlier loopback-compatible mock proof.

vincentkoc · 2026-06-28T17:53:23Z

Merged via squash.

Prepared head SHA: 4dd1874b04f11c88040b07a0228a80b220481b14
Landed commit: a10add753189deaf6a6eb1084f4e2b27786a9660

* fix(models): mark local Ollama rows available * fix(models): mark local Ollama rows available --------- Co-authored-by: Vincent Koc <[email protected]>

* fix(models): mark local Ollama rows available * fix(models): mark local Ollama rows available --------- Co-authored-by: Vincent Koc <[email protected]> (cherry picked from commit a10add7)

* fix(models): mark local Ollama rows available * fix(models): mark local Ollama rows available --------- Co-authored-by: Vincent Koc <[email protected]>

fix(models): mark local Ollama rows available

b5390f5

openclaw-barnacle Bot added commands Command implementations size: XS labels Jun 28, 2026

vincentkoc self-assigned this Jun 28, 2026

fix(models): mark local Ollama rows available

4dd1874

openclaw-barnacle Bot added size: XL and removed size: XS labels Jun 28, 2026

vincentkoc merged commit a10add7 into openclaw:main Jun 28, 2026
182 of 189 checks passed

github-actions Bot mentioned this pull request Jun 28, 2026

📡 Upstream Digest — 2026-06-28 19:05 UTC curtismercier/openclaw-mods#1168

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(models): mark local Ollama rows available#97491

fix(models): mark local Ollama rows available#97491
vincentkoc merged 2 commits into
openclaw:mainfrom
qingminglong:codex/92224-ollama-local-available

qingminglong commented Jun 28, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented Jun 28, 2026 •

edited

Loading

Uh oh!

qingminglong commented Jun 28, 2026

Uh oh!

clawsweeper Bot commented Jun 28, 2026 •

edited

Loading

Uh oh!

qingminglong commented Jun 28, 2026

Uh oh!

Uh oh!

vincentkoc commented Jun 28, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

qingminglong commented Jun 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What Problem This Solves

Why This Change Was Made

User Impact

Evidence

Uh oh!

clawsweeper Bot commented Jun 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

qingminglong commented Jun 28, 2026

Uh oh!

clawsweeper Bot commented Jun 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

qingminglong commented Jun 28, 2026

Uh oh!

Uh oh!

vincentkoc commented Jun 28, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

qingminglong commented Jun 28, 2026 •

edited

Loading

clawsweeper Bot commented Jun 28, 2026 •

edited

Loading

clawsweeper Bot commented Jun 28, 2026 •

edited

Loading