Launch Agent — AI skill for Claude Code
Analyze a task and launch the optimal agent with the right model, effort, and tools.
How to install Launch Agent
Installs to ~/.claude/commands/ciscoittech-claude-agent-framework-launch-agent.md
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/ciscoittech/claude-agent-framework/HEAD/.claude/commands/launch-agent.md -o ~/.claude/commands/ciscoittech-claude-agent-framework-launch-agent.md Restart Claude Code, or start a new session, for it to be picked up.
What Launch Agent does
Analyze a task and launch the optimal agent with the right model, effort, and tools.
Alternatives in AI
- Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
- Open Multi Agent — TypeScript multi-agent orchestration engine — one runTeam() call from goal to result 5.8k ★
- Launch An AI Feature — Workflow recipe — take an AI/LLM feature from a probabilistic-aware PRD to a launch-ready model card by chaini 1.3k ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Smart Task
Multiuse task command that routes work through a JEV-style triage step into the right effort lane and model ti
Prompt Optimize
One-shot prompt rewrite — diagnose a rough draft, steer you to the right workflow archetype (linear, TDD, suba
Goal Agent
Dispatch ONE of the DTD-declared agents on a task — agent name validated against the trust contract, model and
CCW Coordinator Command
Interactive orchestration tool: analyze task → discover commands → recommend chain → execute sequentially → tr
Klaatcode
Open-source AI coding agent for the terminal. Claude Code-grade accuracy with smart model routing — uses the r
Crew Task
Delegate a task to the CodeCrew agent team - auto-routes to the right agents with smart model selection. Use -
Related Agents
Genius Fable
Tier-1 genius worker on Fable. Use for architecture decisions, subtle bugs, algorithm and data-model design, s
Colab Qwen Worker
One-shot task worker - a Qwen model (default qwen3.8:27b via Ollama) on a Google Colab A100 that is started fo
Effort Grader
Blind grader for effortmining. Grades one artifact against a task prompt and a fixed rubric ONLY. The input pa