fix(hermes): route through cliproxy only, drop broken fallback chain - #2335
Conversation
Hermes had its own model-by-model fallback chain (deepseek-pro -> gemma -> glm -> minimax -> free) that never trusted cliproxy's internal quota-exceeded failover. When the primary upstream quota hit, hermes tried each model and 'free' doesn't exist, causing a hard 400 'unknown provider' failure. Point hermes at cliproxy/deepseek-v4-flash as the single default and set fallback_providers to [] so cliproxy handles routing/fallback internally.
|
Warning Indent Zero is shutting down on August 7th. Please migrate over to Indent 2.0 to continue getting PR reviews.
|
|
Warning Review limit reached
Next review available in: 51 minutes You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (2)
📝 WalkthroughSummary by CodeRabbit
WalkthroughBoth Hermes configuration templates now use the Cliproxy DeepSeek Flash endpoint with explicit connection settings. Fallback providers and webhook configuration were removed. The default MOA model reference now uses DeepSeek Flash. ChangesHermes configuration
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches 💡 1🛠️ Fix failing CI checks 💡
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![indent-zero[bot]](https://avatars.githubusercontent.com/in/1426746?s=60&v=4){kind=small}
| model: | ||
| default: cliproxy/deepseek-v4-flash | ||
| base_url: https://cliproxy.shunkakinoki.com/v1 | ||
| api_key: __CLIPROXY_API_KEY__ |
There was a problem hiding this comment.
Redundant credentials on the top-level model: block
providers.cliproxy at lines 5-9 already sets base_url/api_key/api_mode, and custom_providers at the tail sets them a third time. Because model.default uses the prefixed cliproxy/deepseek-v4-flash form, routing resolves through the registered provider without needing credentials here.
On main the analogous delegation: section uses base_url: ''/api_key: '' empty strings, suggesting per-section credentials are optional overrides rather than required. If these top-level fields are genuinely needed to work around a bug, add a comment explaining why; otherwise drop them (in both config.template.yaml and config.tpl.yaml) so there's one source of truth for the cliproxy endpoint/key.
…claw Drop the 'custom/' default prefix (unregistered provider) back to cliproxy/deepseek-v4-flash, but retain the fallback_providers chain (deepseek-v4-pro, gemma, glm, minimax, free) to mirror OpenClaw's model fallbacks, and restore the webhook platform block.
… routing Final minimal change: hermes routes through cliproxyapi (default cliproxy/deepseek-v4-flash) while retaining the fallback_providers chain in sync with OpenClaw (deepseek-v4-pro, gemma, glm, minimax, free) and the webhook platform block.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@config/hermes/config.template.yaml`:
- Around line 2-4: Update model configuration in
config/hermes/config.template.yaml lines 2-4 and config/hermes/config.tpl.yaml
lines 2-4: set model.default to only the model key and add model.provider:
cliproxy. Preserve the existing base URL and API key placeholders, and ensure
MoA and delegation provider/model fields remain explicitly split after
placeholder substitution in both templates.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: cccb9bed-2d87-441d-a67b-50e26781ba65
📒 Files selected for processing (2)
config/hermes/config.template.yamlconfig/hermes/config.tpl.yaml
1 participant
What
Hermes was configured with its own model-by-model fallback chain (deepseek-pro → gemma → glm → minimax →
free) that never trusted cliproxy's internal quota-exceeded failover. When the primary upstream quota hit, hermes tried each model — andfreedoesn't exist, causing a hard 400 'unknown provider' failure.Fix
cliproxy/deepseek-v4-flashas the single default (wascustom/, an unregistered provider prefix)fallback_providers: []so cliproxy handles routing + internal failoverconfig.template.yamlviascripts/llm-update.shTest
deepseek-v4-flashreturns success (routes to Aliyun upstream with quota)Summary by cubic
Route Hermes through
cliproxywithdeepseek-v4-flashas the default, keep fallbacks aligned with OpenClaw, and restore the webhook platform to avoid 400 "unknown provider" errors.base_urlandapi_keyforcliproxy; default model iscliproxy/deepseek-v4-flash.custom/default prefix.webhookplatform (port 8644,__WEBHOOK_SECRET__) and addedcliproxyundercustom_providerswithapi_mode: chat_completions.Written for commit 7651719. Summary will update on new commits.