Repository navigation
Description
Bug type
Behavior bug (incorrect output/state without crash)
Beta release blocker
No
Summary
The manual override to google-vertex model providers through openclaw.json doesn't work
Steps to reproduce
- Add a custom model for google-vertex in openclaw.json
- Restart the gateway
- See the list of registered models for google-vertex
Expected behavior
It should accept the user overrides from openclaw.json and show the additional models in the model list
Actual behavior
openclaw models list --provider google-vertex --all
OpenClaw 2026.6.10 (aa69b12) — If you can describe it, I can probably automate it—or at least make it funnier.
Model Input Ctx Local Auth Tags
google-vertex/gemini-2.5-flash text+image 1024k no yes fallback#1,configured,alias:Gemini Flash
google-vertex/gemini-2.5-flash-lite text+image 1024k no yes
google-vertex/gemini-2.5-pro text+image 1024k no yes default,configured,alias:Gemini Pro
google-vertex/gemini-3-flash-preview text+image 1024k no yes
google-vertex/gemini-3.1-flash-lite text+image 1024k no yes
google-vertex/gemini-3.1-pro-preview text+image 1024k no yes
OpenClaw version
2026.6.10
Operating system
Ubuntu 26.04
Install method
npm global
Model
google-vertex
Provider / routing chain
openclaw -> google-vertex
Additional provider/model setup details
Config for openclaw.json
"models": {
"mode": "merge",
"providers": {
"google-vertex": {
"api": "google-vertex",
"models": [
{
"id": "gemini-3.5-flash",
"name": "Gemini 3.5 Flash (Vertex AI)"
}
]
}
}
},
Logs, screenshots, and evidence
Impact and severity
Affected: anyone using google-vertex models
Severity: Breaks any custom or newer models not in the catalog
Frequency: Always for custom models
Consequence: Limit to static models only
Additional information
No response
Metadata
Metadata
Assignees
Labels
Type
Fields
Priority
Activity
clawsweeper commentedon Jun 24, 2026
Thanks for the context here. I did a careful shell check against current
main, and this is already implemented.Close: current main now covers the reported Google Vertex override path well enough through configured provider row listing, Vertex ADC catalog auth preservation, and a static catalog row for the exact reported model; the broader live Vertex discovery request remains separate.
I found the merged PR that appears to have closed this: #97123: feat(codex): add always plugin approval mode.
So I’m closing this as already implemented rather than keeping a duplicate issue open.
Review details
Best possible solution:
Keep the current main configured-provider and Vertex ADC catalog behavior, and leave live dynamic Vertex discovery to the separate product discussion at #96597.
Do we have a high-confidence way to reproduce the issue?
Yes. The v2026.6.10 report gives concrete
openclaw models list --provider google-vertex --alloutput, and current source shows the corresponding configured-provider and ADC catalog-auth paths that now handle it.Is this the best way to solve the issue?
Yes. Reusing the existing
models.providers.*.modelsand provider catalog/auth normalization paths is the narrow maintainable fix; adding another override surface would duplicate the documented config contract.Security review:
Security review: Non-PR issue triage; no patch security review applies.
AGENTS.md: found and applied where relevant.
What I checked:
--allmodel lists append configuredmodels.providers.*.modelsrows, and rows are eligible when the provider or model declares an API, matching the reporter's provider-levelapiconfig shape. (src/commands/models/list.rows.ts:354, c5d34c8376f8)appendConfiguredProviderRowsafter catalog rows, so arbitrary non-catalog provider models are added when not already seen. (src/commands/models/list.row-sources.ts:79, c5d34c8376f8)mergeProvidersstarts from implicit provider catalog entries and merges explicit provider models back into the provider, preserving the built-in Vertex base URL when user config only adds model rows. (src/agents/models-config.merge.ts:129, c5d34c8376f8)apiKeyfrom env, ADC auth evidence, or auth profiles beforefilterWritableProvidersrequiresbaseUrlandapiKeyfor providers with model rows. (src/agents/models-config.providers.secret-helpers.ts:312, c5d34c8376f8)gemini-3.5-flash, so the exact model in the report no longer depends on manual override to appear. (extensions/google/provider-catalog.ts:39, c5d34c8376f8)src/agents/models-config.applies-config-env-vars.test.ts:470, c5d34c8376f8)Likely related people:
fix: include configured provider rows in all models list, the model-list behavior central to this issue. (role: model-list feature introducer; confidence: high; commits: 5ad02bb32fb6, 2e7ba5412d43; files:src/commands/models/list.rows.ts,src/commands/models/list.list-command.ts)src/agents/models-config.providers.secret-helpers.ts,src/agents/models-config.applies-config-env-vars.test.ts,src/agents/models-config.providers.implicit.ts)gemini-3.5-flashrow to the shared Google/Vertex catalog. (role: recent Google catalog contributor; confidence: high; commits: cd043d270e29, 4d636335dbd1; files:extensions/google/provider-catalog.ts,extensions/google/provider-models.ts)Codex review notes: model internal, reasoning high; reviewed against c5d34c8376f8; fix evidence: merged PR #97123, commit c5d34c8376f8, main fix timestamp 2026-06-27T01:19:00-07:00.