feat(web-search): add explicit provider fallback support by scriptsth · Pull Request #63571 · openclaw/openclaw

scriptsth · 2026-04-09T05:46:09Z

When a provider is explicitly configured with a fallbacks array, retry the next configured provider on any error instead of failing immediately. Without fallbacks, explicit mode keeps existing fail-fast behavior.

Summary

Describe the problem and fix in 2–5 bullets:

If this PR fixes a plugin beta-release blocker, title it fix(<plugin-id>): beta blocker - <summary> and link the matching Beta blocker: <plugin-name> - <summary> issue labeled beta-blocker. Contributors cannot label PRs, so the title is the PR-side signal for maintainers and automation.

Problem:
Why it matters:
What changed:
What did NOT change (scope boundary):

Change Type (select all)

Scope (select all touched areas)

Linked Issue/PR

Closes #
Related #
This PR fixes a bug or regression

Root Cause (if applicable)

For bug fixes or regressions, explain why this happened, not just what changed. Otherwise write N/A. If the cause is unclear, write Unknown.

Root cause:
Missing detection / guardrail:
Contributing context (if known):

Regression Test Plan (if applicable)

For bug fixes or regressions, name the smallest reliable test coverage that should catch this. Otherwise write N/A.

Coverage level that should have caught this:
- Unit test
- Seam / integration test
- End-to-end test
- Existing coverage already sufficient
Target test or file:
Scenario the test should lock in:
Why this is the smallest reliable guardrail:
Existing test that already covers this (if any):
If no new test is added, why not:

User-visible / Behavior Changes

List user-visible changes (including defaults/config).
If none, write None.

Diagram (if applicable)

For UI changes or non-trivial logic flows, include a small ASCII diagram reviewers can scan quickly. Otherwise write N/A.

Before:
[user action] -> [old state]

After:
[user action] -> [new state] -> [result]

Security Impact (required)

New permissions/capabilities? (Yes/No)
Secrets/tokens handling changed? (Yes/No)
New/changed network calls? (Yes/No)
Command/tool execution surface changed? (Yes/No)
Data access scope changed? (Yes/No)
If any Yes, explain risk + mitigation:

Repro + Verification

Environment

OS:
Runtime/container:
Model/provider:
Integration/channel (if any):
Relevant config (redacted):

Steps

Configure tools.web.search.provider = "google" with tools.web.search.fallbacks = ["duckduckgo", "bing", "searx"]
Make a web search request that triggers a provider error (e.g., invalid API key, network failure)
Observe that the search retries with each fallback provider in order (google → duckduckgo → bing → searx) until one succeeds
Verify the final result comes from the first available fallback provider

Expected

fallbacks logic worked as expected

Actual

fallbacks logic worked as expected

Evidence

Attach at least one:

Failing test/log before + passing after
Trace/log snippets
Screenshot/recording
Perf numbers (if relevant)

Human Verification (required)

What you personally verified (not just CI), and how:

Verified scenarios:
Edge cases checked:
What you did not verify:

Review Conversations

I replied to or resolved every bot review conversation I addressed in this PR.
I left unresolved only the conversations that still need reviewer or maintainer judgment.

If a bot review conversation is addressed by this PR, resolve that conversation yourself. Do not leave bot review conversation cleanup for maintainers.

Compatibility / Migration

Backward compatible? (Yes)
Config/env changes? (Yes)
Migration needed? (No)
If yes, exact upgrade steps:

Risks and Mitigations

List only real risks for this PR. Add/remove entries as needed. If none, write None.

Risk:
- None

When a provider is explicitly configured with a fallbacks array, retry the next configured provider on any error instead of failing immediately. Without fallbacks, explicit mode keeps existing fail-fast behavior. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f64b43f0b1

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

greptile-apps · 2026-04-09T05:51:34Z

Greptile Summary

Adds explicit provider fallback support to runWebSearch: when tools.web.search.fallbacks is configured alongside an explicit provider, the runtime retries each listed fallback in order on failure; without fallbacks, explicit-provider mode keeps the existing fail-fast behavior. Config type, Zod schema, generated schema, labels, help text, and baseline hash are all updated consistently.

The duplicate import declarations for resolvePluginWebSearchProviders and resolveRuntimeWebSearchProviders from the same module path (../plugins/web-search-providers.runtime.js) at lines 7–8 of runtime.ts will fail the eslint/no-duplicate-imports correctness rule enabled by categories.correctness: \"error\" in .oxlintrc.json; they should be merged into one statement before merge.

Confidence Score: 4/5

Safe to merge after fixing the duplicate import lint failure.

The fallback logic itself is correct, well-tested, and the config surface is consistently updated. Score is 4 rather than 5 because the duplicate import declarations from the same module path will cause pnpm check to fail under the repo's correctness: error oxlint policy, which is a required gate before landing on main.

src/web-search/runtime.ts (duplicate import at lines 7–8)

Vulnerabilities

No security concerns identified. The change adds provider-retry logic that stays within the existing credential and provider-registry surfaces — no new network endpoints, no secrets handling changes, and no expansion of access scope.

Comments Outside Diff (1)

src/web-search/runtime.ts, line 7-8 (link)

Duplicate import declarations from the same module

Both resolvePluginWebSearchProviders and resolveRuntimeWebSearchProviders are imported from ../plugins/web-search-providers.runtime.js in separate statements. With categories.correctness: "error" in .oxlintrc.json, the eslint/no-duplicate-imports rule will fire on this and cause pnpm check to fail.

Prompt To Fix With AI

This is a comment left during a code review.
Path: src/web-search/runtime.ts
Line: 7-8

Comment:
**Duplicate import declarations from the same module**

Both `resolvePluginWebSearchProviders` and `resolveRuntimeWebSearchProviders` are imported from `../plugins/web-search-providers.runtime.js` in separate statements. With `categories.correctness: "error"` in `.oxlintrc.json`, the `eslint/no-duplicate-imports` rule will fire on this and cause `pnpm check` to fail.



How can I resolve this? If you propose a fix, please make it concise.

Prompt To Fix All With AI

This is a comment left during a code review.
Path: src/web-search/runtime.ts
Line: 7-8

Comment:
**Duplicate import declarations from the same module**

Both `resolvePluginWebSearchProviders` and `resolveRuntimeWebSearchProviders` are imported from `../plugins/web-search-providers.runtime.js` in separate statements. With `categories.correctness: "error"` in `.oxlintrc.json`, the `eslint/no-duplicate-imports` rule will fire on this and cause `pnpm check` to fail.

```suggestion
import {
  resolvePluginWebSearchProviders,
  resolveRuntimeWebSearchProviders,
} from "../plugins/web-search-providers.runtime.js";
```

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: src/web-search/runtime.ts
Line: 333-337

Comment:
**Redundant optional chaining in type-predicate filter**

`Boolean(p)` already guards against `null`/`undefined` before the `&&`, so `p?.id` on the next condition can never be nullish — the `?.` is redundant. Using plain `p.id` here keeps the intent clearer and avoids a misleading suggestion that `p` could still be absent at that point.

```suggestion
          .filter(
            (p): p is PluginWebSearchProviderEntry => Boolean(p) && p.id !== primaryProviderId,
          ),
```

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (1): Last reviewed commit: "feat(web-search): add explicit provider ..." | Re-trigger Greptile}

Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1542f95790

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Use normalizeLowercaseStringOrEmpty on fallback ids to ensure consistent matching with primary provider field, handling mixed-case/whitespace. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 948ce231bb

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

When providerId is explicitly provided (for example via CLI --provider override), ignore config fallbacks and fail fast to preserve deterministic provider-specific runs and debugging. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 000f534310

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Prevent the same provider from being retried multiple times when duplicate fallback IDs are configured, avoiding extra latency and repeated billable requests on failure paths. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 938bb54f43

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 938bb54f43

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

When explicit fallback mode is enabled, resolve the primary provider from the config provider field instead of relying on runtime-ordered candidates, preventing stale runtime metadata from bypassing the intended primary. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 348b0b21ee

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

…lves Previously allowFallback was enabled whenever tools.web.search.fallbacks was non-empty, even if every fallback id was invalid. Now it only enables fallback when at least one usable fallback candidate exists. Co-Authored-By: Claude Opus 4.6 <[email protected]>

scriptsth · 2026-04-09T07:36:43Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5041018ad1

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

…anning When tools.web.search.provider is set to an unknown id but runtime metadata marks selection as explicit, the fallback path is now skipped and candidates are used instead, preventing Error("undefined") when no fallback resolves. Co-Authored-By: Claude Opus 4.6 <[email protected]>

scriptsth · 2026-04-09T11:46:12Z

@codex review

chatgpt-codex-connector

💡 Codex Review

openclaw/src/web-search/runtime.ts

Lines 385 to 388 in df03900

    
           return { 
        
             provider: candidate.id, 
        
             result: await definition.execute(params.args), 
        
           };

Treat provider error payloads as fallback failures

This fallback path only retries on thrown exceptions, but many web-search providers report auth/validation failures by returning an { error: ..., message: ... } payload instead of throwing (for example, missing-key responses in provider implementations). In those cases runWebSearch returns the primary provider response immediately and never reaches configured tools.web.search.fallbacks, so explicit fallback mode silently fails for common real-world failures like missing or invalid credentials.

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Many providers return { error: ... } payloads instead of throwing on auth failures. Now these are treated as failures and trigger fallback retry, not immediate return. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 48136ac7b4

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

…failure Return error payloads instead of throwing to preserve structured fields like message/docs that providers return for user-facing diagnostics. Co-Authored-By: Claude Opus 4.6 <[email protected]>

scriptsth · 2026-04-09T12:52:34Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9aa16cf436

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Only retry error payloads when explicit fallback mode is configured, preserving existing auto-detect behavior. Co-Authored-By: Claude Opus 4.6 <[email protected]>

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 706c365926

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

steipete · 2026-07-09T14:46:49Z

Thanks @scriptsth. Closing this stale conflicting implementation branch while keeping #80843 open. This head retries every provider error instead of only retryable failures and can silently drop configured fallback provider IDs, with no live fallback proof. A replacement should start from current main, preserve the full fallback chain, and test retry classification explicitly.

openclaw-barnacle Bot added docs Improvements or additions to documentation size: M labels Apr 9, 2026

chatgpt-codex-connector Bot reviewed Apr 9, 2026

View reviewed changes