[Bug]: startup readiness error text over-attributes network failures to provider/API settings

[zhyongrui]

Bug type

Behavior bug (incorrect output/state without crash)

Beta release blocker

No

Summary

When the startup-active readiness probe fails with LLM request failed: network connection error, the user-facing startup message says 请检查 provider/API 设置, which attributes the failure to provider/API configuration even though the observed error is network-related.

Steps to reproduce

1. Start OpenClaw 2026.4.14 with Feishu enabled and the default model set to openai/gpt-5.4.
2. Let the startup-active readiness probe run.
3. Reproduce an LLM call failure where the probe logs LLM request failed: network connection error.
4. Observe the proactive user-facing message sent after startup: OpenClaw 已启动，但当前模型配置不可用，请检查 provider/API 设置。详情：LLM request failed: network connection error.

Expected behavior

The user-facing message should stay within what is directly supported by the observed error. For this case, a message like OpenClaw 已启动，但当前配置不可用。详情：LLM request failed: network connection error. would be accurate. It should not specifically tell the user to check provider/API settings unless the evidence points to an auth or provider configuration problem.

Actual behavior

The startup notification says 请检查 provider/API 设置 even though the observed error in logs is LLM request failed: network connection error.

OpenClaw version

2026.4.14 (323493f)

Operating system

Ubuntu 24.04.3 LTS

Install method

pnpm global

Model

openai/gpt-5.4

Provider / routing chain

openclaw -> local HTTP proxy (127.0.0.1:7890) -> openai

Additional provider/model setup details

Observed config had agents.defaults.model.primary = openai/gpt-5.4. The gateway service environment included HTTP_PROXY=http://127.0.0.1:7890 and HTTPS_PROXY=http://127.0.0.1:7890.

Logs, screenshots, and evidence

User-facing message:
OpenClaw 已启动，但当前模型配置不可用，请检查 provider/API 设置。
详情：LLM request failed: network connection error.

Gateway log excerpts:
2026-04-16T15:59:59.659+00:00 [agent/embedded] embedded run agent end: runId=openclawcode-feishu-readiness-dd3d2ea1-3a8c-42e8-96f9-ac4670a4769d isError=true model=gpt-5.4 provider=openai error=LLM request failed: network connection error. rawError=Connection error.
2026-04-16T16:00:20.132+00:00 [plugins] openclawcode feishu readiness probe failed for startup-active via openai/gpt-5.4: LLM request failed: network connection error.

Impact and severity

Affected: users receiving the startup-active readiness failure notification
Severity: Low to medium; the product still starts, but the message points users at a potentially wrong troubleshooting path
Frequency: observed once during onboarding on this machine
Consequence: misleading guidance during setup and debugging

Additional information

This report is about the wording scope, not the underlying network failure itself. The request here is to make the user-facing text more evidence-based for non-auth/non-provider failures.