hermes-custom-auxiliary-providers
Configure Hermes auxiliary tasks (vision, compression, web_extract) with custom API providers or native built-in providers (Gemini, etc.), plus fallback…
Skill by statertly-pixel
Chalk Labs
Need help deploying hermes-custom-auxiliary-providers?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/statertly-pixel-hermes-custom-auxiliary-providers
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: backend
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/hermes-custom-auxiliary-providers
What's inside
- Skill name
hermes-custom-auxiliary-providers- Description
- Configure Hermes auxiliary tasks (vision, compression, web_extract) with custom API providers or native built-in providers (Gemini, etc.), plus fallback provider chains.
- License
- MIT
- Instructions
- 272 lines in SKILL.md · view on GitHub
Before installing
- Instructions call the network
Files (5)
- SKILL.md
- dashscope-vision-setup.md · references/
- gemini-auxiliary-bulk-setup.md · references/
- gemini-web-extract-setup.md · references/
- siliconflow-vision-setup.md · references/
SKILL.md (first lines)
# Hermes Custom Auxiliary Providers
Configure Hermes auxiliary tasks (vision, compression, web_extract, skills_hub, approval, etc.) with non-standard API providers that aren't built-in.
## When to Use
- Your main model (e.g. DeepSeek) lacks vision/multimodal support and you need a separate vision API
- You want to route different auxiliary tasks (vision, compression, summarization) to different cheaper/faster models
- You need to offload web extraction (GitHub, YouTube, Google pages) to a model better suited for external content, keeping your main model for reasoning
- You're using a China-based provider like SiliconFlow, DashScope (Alibaba), Zhipu, or MiniMax for auxiliary tasks
- You have a custom/self-hosted endpoint that serves an OpenAI-compatible API
- You want to use a native built-in provider (like `gemini` or `anthropic`) for auxiliary tasks instead of a `custom` endpoint
## Procedure
### 1. Identify the auxiliary task to configure
Common auxiliary tasks:
- `auxiliary.vision` — image analysis (vision_analyze tool)
- `auxiliary.compression` — context compression
- `auxiliary.web_extract` — web page content extraction
- `auxiliary.skills_hub` — skill hub searches
- `auxiliary.approval` — smart command approval
### 2. Configure via CLI
```bash
# Set provider to "custom" (tells Hermes to use the explicit base_url + api_key)
hermes config set auxiliary.vision.provider custom
# Set the model name as the provider's API expects it
hermes config set auxiliary.vision.model Qwen/Qwen3-Omni-30B-A3B-Captioner
# Set the API base URL (OpenAI-compatible endpoint)
hermes config set auxiliary.vision.base_url https://api.siliconflow.cn/v1
# Set the API key (stored in config.yaml)
hermes config set auxiliary.vision.api_key 'your-api-key-here'
```
The same pattern works for other auxiliary tasks:
```bash
hermes config set auxiliary.compression.provider custom
hermes config set auxiliary.compression.model your-model-name
hermes config set auxiliary.compression.base_url https://your-api-endpoint/v1
hermes config set auxiliary.compression.api_key 'your-api-key'
```
### 3. Discover available models (optional)
List models on the provider to find the right model ID:
```bash
# SiliconFlow
curl -s https://api.siliconflow.cn/v1/models | python3 -c "import sys,json;[print(m['id']) for m in json.load(sys.stdin)['data']]"
# DashScope
curl -s -H "Authorization: Bearer YOUR_KEY" https://dashscope.aliyuncs.com/compatible-mode/v1/models | python3 -c "import sys,json;[print(m['id']) for m in json.load(sys.stdin)['data']]"
```
### 4. Test the connection
Verify the provider works with a direct API call before configuring:
```bash
# For vision: send a base64-encoded image
python3 -c "
import requests, json, base64
key = 'your-api-key'
img_path = r'C:\path\to\test.jpg'
with open(img_path, 'rb') as f:
b64 = base64.b64encode(f.read()).decode()
data_url = 'data:image/jpeg;base64,' + b64
r = requests.post('https://api.siliconflow.cn/v1/chat/completions',
headers={'Authorization': 'Bearer ' + key, 'Content-Type': 'application/json'},
json={'model': 'Qwen/Qwen3-Omni-30B-A3B-Captioner',
'messages': [{'role': 'user', 'content': [
{'type': 'text', 'text': 'Describe this image'},
{'type': 'image_url', 'image_url': {'url': data_url}}
]}], 'max_tokens': 200}, timeout=60)
print(r.json()['choices'][0]['message']['content'] if r.ok else r.text)
"
```
### 5. Restart session
Config changes only take effect on session restart:
- CLI: exit and relaunch `hermes`, or `/reset`
- Gateway: `/restart`
## Common Custom Providers
### SiliconFlow (硅基流动)
- Base URL: `https://api.siliconflow.cn/v1`
- OpenAI-compatible chat completions API
- Supports Qwen, DeepSeek, GLM, MiniMax, and many other models
- API key access varies per model (some require Pro tier)
- Tested working models for vision: `Qwen/Qwen3-Omni-30B-A3B-Captioner`
- Models that require higher-tier key: `deepseek-ai/DeepSeek-OCR`, `nex-agi/Nex-N2-Pro`, `PaddlePaddle/PaddleOCR-VL-1.5`
### Alibaba DashScope (阿里云百炼)
- Base URL: `https://dashscope.aliyuncs.com/compatible-mode/v1`
- OpenAI-compatible chat completions API
- **Cheapest vision model**: `qwen3-vl-flash` (~¥0.003/1k input tokens) — fast, decent quality for general image description
- Mid-range: `qwen3-vl-plus` (~¥0.01/1k tokens) — better quality, slightly more expensive
- Strongest: `qwen-vl-max` — most capable but most expensive
- OCR-focused: `qwen-vl-ocr` — specialized for text extraction
- Key format: `sk-` prefix, typically ~116 chars; supports Bearer auth
- API key must be tested with `POST /v1/chat/completions`, not just `GET /v1/models`
## Available Vision Models (as of 2026-06)
**SiliconFlow:**
- `Qwen/Qwen3-Omni-30B-A3B-Captioner` (general captioning, works with standard key)
- `deepseek-ai/DeepSeek-OCR` (OCR only, requires Pro tier)
- `nex-agi/Nex-N2-Pro` (requires Pro tier)
- `PaddlePaddle/PaddleOCR-VL-1.5` (requires Pro tier)
**DashScope:**From statertly-pixel/mcp-cloud-deploy, MIT.
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jul 18, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get hermes-custom-auxiliary-providers running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
hermes-custom-auxiliary-providers skill. Agent Mart. https://agentmart.dev/skills/statertly-pixel-hermes-custom-auxiliary-providers (last updated 2026-10-24)