hermes-model-config

Configure and audit Hermes Agent model providers, fallback chains, and cron overrides.

1|Updated Jun 21, 2026
One-click install
npx skills add https://github.com/tapway/shogun-os --skill hermes-model-config-tapway
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hermes-model-config
Source: https://github.com/tapway/shogun-os/tree/main/skills/hermes/hermes-model-config
Command: npx skills add https://github.com/tapway/shogun-os --skill hermes-model-config-tapway

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Hermes Agent model configuration spans primary providers, named custom providers, and fallback chains across multiple profiles and cron jobs, and misconfigurations like silent fallback traps, truncated API keys, and stale credential pool entries cause hidden billing spikes and HTTP 401/404 failures that are hard to diagnose from config files alone. ## Core Features & Use Cases - Three-Tier Provider Configuration: Set up primary models, named custom_providers (OpenAI or Anthropic wire protocol), and fallback_providers chains in config.yaml with correct YAML structure. - Silent Fallback Detection: Audit agent.log to count real provider failures versus successes, revealing when a broken primary routes 100% of traffic to an expensive fallback like OpenRouter. - Credential & Auxiliary Fixes: Detect truncated API keys containing literal '...', clean stale credential_pool entries in auth.json, and fix HTTP 401 errors on auxiliary features like title generation and compression. - Use Case: A user notices high OpenRouter billing after switching to DashScope. The audit procedure reveals 145 DashScope 404 failures and 3,834 OpenRouter successes in agent.log, pinpointing an endpoint/model-name mismatch as the root cause. ## Quick Start Ask the agent to audit which model provider is actually serving traffic across all Hermes profiles and explain any unexpected fallback usage.

Frequently Asked Questions about hermes-model-config

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I configure a custom model provider in Hermes Agent?

Add a custom_providers entry as a YAML list at the top level of config.yaml with name, base_url, api_key, and provider_type (openai or anthropic). Reference it as provider: custom:<name>, or use provider: custom with inline base_url and api_key under the model block.

How do I detect silent fallback to a backup model provider?

Count failures and successes in agent.log: grep for NotFoundError with provider=custom on the primary and compare against provider=openrouter successes on the fallback. If every primary call fails with HTTP 404 or 401, the fallback chain handles 100% of traffic with no visible error.

Why does Hermes title generation fail with HTTP 401 when chat works?

Auxiliary sections with provider: auto do not carry the inline api_key from model.provider: custom through their code path. Fix by pinning each auxiliary section (title_generation, compression, triage_specifier) to a named custom provider that has its own api_key.

Why is my Hermes API key rejected even though it looks correct?

The key may contain literal '...' from a truncation artifact, making it 13 characters instead of 35. Verify with len() in Python rather than visual inspection, since read_file displays valid 35-char keys with the same '...' pattern.

Can Hermes fallback providers prevent a gateway crash from rate limiting?

No. The fallback_providers chain engages at runtime for API failures but not during gateway initialization. A rate-limited provider (HTTP 429) kills the gateway at boot before fallback logic runs, requiring a manual provider switch and restart.

When should I not use this model configuration skill?

Do not use it for gateway lifecycle management (use hermes-agent), cron job creation (use hermes-cron-jobs), or profile persona authoring (use profile-management). It covers only model provider configuration, auditing, and troubleshooting.