smart-spawn-api

What This Skill Does

The Smart Spawn API skill is a powerful infrastructure tool for OpenClaw that automates model selection for any given task. By integrating directly with the ss.deeflect.com API, it eliminates the guesswork of choosing the right LLM for specific requirements. Whether you are running a high-level coding project, a rapid creative brainstorming session, or a complex research task, this skill analyzes your objective, budget, and performance requirements to recommend the most optimal model ID. You can then pass this model ID directly into your sessions_spawn commands to ensure your agent uses the best possible tool for the job. Beyond simple selection, it supports complex workflows including task decomposition, parallel swarm execution, and model comparisons, making it an essential utility for multi-agent orchestration.

Installation

To integrate this capability into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/deeflect/smart-spawn

Once installed, the skill acts as a middleware layer. You do not need to configure complex plugins; it operates via standard HTTP requests, meaning it is instantly compatible with your existing OpenClaw workflows.

Use Cases

• Efficient Resource Management: Automatically select low-budget models for repetitive, high-volume tasks while reserving high-budget, high-capability models for critical reasoning or architectural work.
• Automated Workflow Orchestration: Use the /api/decompose or /api/swarm endpoints to break large, ambiguous projects into smaller, manageable subtasks, each assigned to a model specifically suited for that step.
• Vendor Agnosticism: Easily swap between models from various providers (Anthropic, OpenAI, etc.) based on real-time performance scores and pricing without needing to manually update configuration files.

Example Prompts

1. "Smart spawn a new coding session for a React dashboard. Use the Smart Spawn API to pick an appropriate medium-budget model."
2. "I have a complex task to build a SaaS app. Can you decompose this into steps using the Smart Spawn API and start the sub-agent swarm?"
3. "Compare the pricing and performance capabilities between Claude Opus 4.6 and GPT-5.2 using the model comparison tool."

Tips & Limitations

• Budget Management: Always define your budget (low, medium, high, or any). If you encounter a 404 error, it usually means the requested constraints are too narrow—use the 'any' parameter to ensure a fallback is always found.
• Error Handling: The skill is designed for robustness. If the external API is unreachable or rate-limited, fall back to your default OpenClaw spawning logic. Avoid hard-coding specific model IDs when possible; rely on the API to provide them dynamically to stay updated with the latest model releases.
• Contextual Awareness: Use the context parameter to provide hints like 'vision' or 'long-context' to significantly improve the accuracy of the API's recommendations for specialized tasks.