Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
v1.5.0: thinking/reasoning mode (--thinking flag, Qwen3 235B Thinking + Kimi K2 Thinking/K2.5), 5 new models (GPT-5.2 Codex, Gemini 3.1 Pro 1M ctx, Hermes 3 405B), fix budget-medium classifier gap, SKILL.md spec polish (YAML metadata, cliHelp, os field)
What This Skill Does
Routes prompts to Venice.ai models, automatically selecting the cheapest model adequate for each query's complexity. Supports zero-retention private mode, uncensored models, web search with citations, function calling, conversation history, and daily or session cost budgets.
Automatically picks the cheapest Venice.ai model capable of handling a given prompt, avoiding manual model selection while keeping all prompts off Big Tech infrastructure.
When to use it
- Analyzing a confidential contract through a zero-retention private model
- Writing creative or edgy fiction routed to uncensored models
- Looking up current news via web search with cited sources
- Capping daily API spend with automatic tier downgrade near budget limits
- Multi-turn code review with context-aware model selection across the conversation
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: --private-only --prompt "Summarize this acquisition agreement for key risk clauses"
- 1Classifies prompt complexity using keyword and length heuristics
- 2Applies private-only mode to restrict candidates to zero-retention Venice.ai models
- 3Selects tier and model appropriate for legal document analysis
- 4Sends the prompt to Venice.ai and streams the response back
Summary of key risk clauses from the acquisition agreement, processed through a private Venice.ai model with zero data retention
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.