fix(status): show configured fallback models in /status output by AnCoSONG · Pull Request #33111 · openclaw/openclaw

AnCoSONG · 2026-03-03T14:36:21Z

Previously, the /status command only showed fallback models when they were actively being used (i.e., when the primary model failed and a fallback was activated). This made it hard for users to verify their fallback configuration without checking the raw config.

Changes

Added "configuredFallbacksLine" to display configured fallback models from "agents.defaults.model.fallbacks"
The fallback line shows as "🔄 Fallbacks: model1, model2" when fallbacks are configured
Active fallback state is still shown separately with "↪️ Fallback: ..."
Added tests for the new functionality

Example Output

With config:
"""json
"model": {
"primary": "anthropic/claude-opus-4-6",
"fallbacks": ["google/gemini-2.5-flash"]
}
"""

Status now shows:
""\

greptile-apps · 2026-03-03T14:39:39Z

Greptile Summary

This PR adds a configuredFallbacksLine to the /status output so users can see which fallback models are configured before a live failover actually happens. The core change is a small read in buildStatusMessage that surfaces agent.model.fallbacks, and a corresponding update in buildStatusText that resolves agent-specific vs. inherited fallbacks via resolveAgentModelFallbacksOverride before forwarding them. Tests cover the primary scenarios: per-agent override, inheritance from defaults, and the empty-array "disable inherited fallbacks" case.

Confidence Score: 5/5

Safe to merge — all remaining findings are P2 or lower and do not affect correctness.

The logic is straightforward: read model.fallbacks from the resolved agent config and format it. The three-tier inheritance (agent override → agent has no override, inherit defaults → empty array disables) is handled correctly and covered by targeted tests. The fallback state display is additive and does not change any existing behavior.

No files require special attention.

Vulnerabilities

No security concerns identified. The change reads model config strings from the already-validated config object and formats them for display only.

_{Reviews (2): Last reviewed commit: "fix(status): keep fallback changelog ent..." | Re-trigger Greptile}

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 39aabbab2a

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 · 2026-03-03T14:40:32Z

src/auto-reply/status.ts

+    const modelConfig = args.agent?.model;
+    if (typeof modelConfig === "object" && modelConfig && Array.isArray(modelConfig.fallbacks)) {
+      return modelConfig.fallbacks;


Derive displayed fallbacks from effective agent settings

Reading modelConfig.fallbacks from args.agent here causes /status to report the wrong fallback list for non-default agents, because both status callers currently construct args.agent from cfg.agents.defaults (src/auto-reply/reply/commands-status.ts:162-170, src/agents/tools/session-status-tool.ts:351-356). If an agent overrides model.fallbacks (or sets [] to disable global fallbacks), the new line will still show default fallbacks, so users get incorrect configuration feedback even though runtime fallback resolution honors per-agent overrides.

Useful? React with 👍 / 👎.

AnCoSONG · 2026-03-04T13:31:44Z

关于 CI 失败的说明

当前 CI 中的失败看起来是已知的 flaky tests，而非本 PR 代码改动造成：

失败的 checks：

checks (bun, test) - Issue [Bug]: Flaky test in bun runner: isolated-agent model override precedence expects 'openai' but gets 'anthropic' #32084 记录的 Bun runner flaky test
checks-windows (node, test, shards 2, 5) - Issue bug: flaky Windows CI test in path-safety (checks-windows) #30878 记录的 Windows CI path-safety flaky test
macos - 超时 2h+，可能是基础设施问题

验证：

✅ 本地 pnpm test：28/28 通过
✅ Linux CI (checks node test)：通过

建议维护者可以重跑失败的 checks，或者考虑在已知 flaky 的情况下 review 合并。

cc @maintainers

altaywtf · 2026-04-09T08:50:39Z

@greptileai review

Copilot

Pull request overview

Updates the /status command (and shared status card output) to always display the configured model fallback chain, not only the active fallback when a failover happens. This improves visibility into agents.defaults.model.fallbacks and per-agent fallback overrides.

Changes:

Add a new “🔄 Fallbacks: …” line in buildStatusMessage() when fallbacks are configured.
Wire per-agent fallback overrides into /status construction so the displayed list matches the effective agent configuration.
Add/extend tests to cover configured fallbacks display, inheritance, and explicit disabling.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
src/auto-reply/status.ts	Adds rendering of configured fallback models into the status message output.
src/auto-reply/status.test.ts	Adds unit tests asserting the configured fallbacks line appears/omits correctly.
src/auto-reply/reply/commands-status.ts	Threads per-agent fallback override into the status message’s agent model config.
src/auto-reply/reply/commands-status.thinking-default.test.ts	Adds integration-style tests verifying per-agent fallback overrides/inheritance in status reply output.
CHANGELOG.md	Documents the behavior change for `/status` and shared status cards.

Copilot · 2026-04-09T08:56:03Z

src/auto-reply/status.ts

+  // Show configured fallback models (from agent model config)
+  const configuredFallbacks = (() => {
+    const modelConfig = args.agent?.model;
+    if (typeof modelConfig === "object" && modelConfig && Array.isArray(modelConfig.fallbacks)) {
+      return modelConfig.fallbacks;
+    }
+    return undefined;
+  })();
+  const configuredFallbacksLine = configuredFallbacks?.length
+    ? `🔄 Fallbacks: ${configuredFallbacks.join(", ")}`
+    : null;


The configured fallback list is displayed verbatim via configuredFallbacks.join(", "). Since AgentModelSchema allows arbitrary strings (including whitespace/empty strings and duplicates), /status can show blank entries or repeated models (e.g., Fallbacks: , openai/gpt-4.1, openai/gpt-4.1). Consider normalizing the list before rendering (trim, drop empty, de-dupe case-insensitively) to keep status output clean and consistent with other places that normalize fallback lists (e.g., gateway agent listing).

Fixes openclaw#33099 Previously, the /status command only showed fallback models when they were actively being used. Now it also displays configured fallback models from agents.defaults.model.fallbacks so users can verify their setup at a glance. - Added configuredFallbacksLine to show fallback models from config - Added tests for the new functionality

PR Review Feedback: - Use resolveAgentModelFallbacksOverride() to get agent-specific fallbacks - Fixes incorrect fallback display for non-default agents - Agent can now set fallbacks: [] to disable global fallbacks

altaywtf · 2026-04-09T09:13:15Z

Merged via squash.

Prepared head SHA: 5e590aa68c734b0c513bc2f6935305a3766b4760
Merge commit: 1b24560392528ce84ba82b6edab753eda9bb0a7f

Thanks @AnCoSONG!

@altaywtf

Merged via squash. Prepared head SHA: 5e590aa Co-authored-by: AnCoSONG <32268203+AnCoSONG@users.noreply.github.com> Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com> Reviewed-by: @altaywtf

openclaw-barnacle bot added the size: XS label Mar 3, 2026

chatgpt-codex-connector bot reviewed Mar 3, 2026

View reviewed changes

openclaw-barnacle bot added agents Agent runtime and tooling size: S and removed size: XS labels Mar 3, 2026

altaywtf self-assigned this Apr 8, 2026

altaywtf force-pushed the fix/status-show-fallback-models branch from f86fe90 to 35a76cc Compare April 8, 2026 10:58

openclaw-barnacle bot removed the agents Agent runtime and tooling label Apr 8, 2026

altaywtf force-pushed the fix/status-show-fallback-models branch 7 times, most recently from dbef32f to 8e15c41 Compare April 9, 2026 08:49

altaywtf requested a review from Copilot April 9, 2026 08:50

Copilot started reviewing on behalf of altaywtf April 9, 2026 08:51 View session

Copilot AI reviewed Apr 9, 2026

View reviewed changes

AnCoSONG and others added 6 commits April 9, 2026 10:12

fix(status): derive fallbacks from effective agent settings

3991da3

PR Review Feedback: - Use resolveAgentModelFallbacksOverride() to get agent-specific fallbacks - Fixes incorrect fallback display for non-default agents - Agent can now set fallbacks: [] to disable global fallbacks

ci: trigger rebuild

0dd4dd6

fix(status): show configured fallback models

5566a5c

fix(status): preserve inherited fallback visibility

4dba520

fix(status): keep fallback changelog entry unreleased

5e590aa

altaywtf force-pushed the fix/status-show-fallback-models branch from 8e15c41 to 5e590aa Compare April 9, 2026 09:12

altaywtf merged commit 1b24560 into openclaw:main Apr 9, 2026
24 of 26 checks passed

github-actions bot mentioned this pull request Apr 9, 2026

📡 Upstream Digest — 2026-04-09 10:53 UTC curtismercier/openclaw-mods#519

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(status): show configured fallback models in /status output#33111

fix(status): show configured fallback models in /status output#33111
altaywtf merged 6 commits intoopenclaw:mainfrom
AnCoSONG:fix/status-show-fallback-models

AnCoSONG commented Mar 3, 2026

Uh oh!

greptile-apps bot commented Mar 3, 2026 •

edited

Loading

Vulnerabilities

Uh oh!

chatgpt-codex-connector bot left a comment

Uh oh!

chatgpt-codex-connector bot Mar 3, 2026

Uh oh!

AnCoSONG commented Mar 4, 2026

Uh oh!

altaywtf commented Apr 9, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 9, 2026

Uh oh!

Uh oh!

altaywtf commented Apr 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

AnCoSONG commented Mar 3, 2026

Changes

Example Output

Uh oh!

greptile-apps bot commented Mar 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Greptile Summary

Confidence Score: 5/5

Vulnerabilities

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector bot Mar 3, 2026

Choose a reason for hiding this comment

Uh oh!

AnCoSONG commented Mar 4, 2026

Uh oh!

altaywtf commented Apr 9, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

altaywtf commented Apr 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

greptile-apps bot commented Mar 3, 2026 •

edited

Loading