Model Route Guard

Diagnose and fix model routing conflicts. Ensure primary model uses correct provider endpoint without duplicate overrides.

Fix model routing and provider endpoint conflicts.

Problem

Model routing issues cause:

• Wrong provider endpoint used
• Duplicate provider definitions
• Agent overrides conflicting with global config
• Silent fallback to wrong model

Workflow

1. Route Audit

powershell

# Check global config
$cfg = Get-Content "$HOME/.openclaw/openclaw.json" -Raw | ConvertFrom-Json
$globalBase = $cfg.models.providers.bailian.baseUrl
$globalModel = $cfg.agents.defaults.model.primary

# Check agent overrides
$agentCfgPath = "$HOME/.openclaw/agents/main/agent/models.json"
if (Test-Path $agentCfgPath) {
    $agentCfg = Get-Content $agentCfgPath -Raw | ConvertFrom-Json
    $agentBase = $agentCfg.providers.bailian.baseUrl
}

"Global baseUrl = $globalBase"
"Global model   = $globalModel"
"Agent baseUrl  = $agentBase"

# Detect conflicts
if ($globalBase -ne $agentBase) {
    Write-Warning "Provider URL mismatch between global and agent config"
}

2. Fix Conflicts

powershell

# Correct endpoint (coding.dashscope, not coding-intl)
$correctUrl = "https://coding.dashscope.aliyuncs.com/v1"

# Update global config
$cfg.models.providers.bailian.baseUrl = $correctUrl
$cfg | ConvertTo-Json -Depth 10 | Out-File "$HOME/.openclaw/openclaw.json" -Encoding UTF8

# Remove conflicting agent override
if (Test-Path $agentCfgPath) {
    Remove-Item $agentCfgPath -Force
}

# Restart
openclaw gateway restart

3. Verification

powershell

# Test model call
openclaw models list

# Check active route
openclaw status

Executable Completion Criteria

Criteria	Verification
Single provider URL	Only one bailian.baseUrl in config
Correct endpoint	URL is coding.dashscope.aliyuncs.com
No duplicate overrides	models.json removed or aligned
Model call succeeds	openclaw models list works

Privacy/Safety

• No API keys in audit output
• Redact key values (show first 4 chars)
• Config files only, no external logging

Self-Use Trigger

Use when:

• Model calls fail with endpoint errors
• Config shows multiple provider definitions
• After gateway restart with model issues
• Silent fallback to wrong model detected

---

One route. One source of truth.