TMA1 v2: a workflow pattern to sync multiple Claude agents without conflict
A community member shared 'TMA1 v2', a workflow design that lets multiple AI agents share context and stay in sync while working on the same project. It combines Claude Code's Hooks, Skills, and MCP to coordinate agents without them stepping on each other. It's a hands-on reference for anyone building multi-agent systems with Claude.
When you run several AI agents at the same time, they often end up doing the same work twice or conflicting because each one only knows what it has seen. TMA1 v2 tackles this by wiring together three Claude Code features: Hooks (commands that fire automatically before or after an action), Skills (reusable capability modules you call by name), and MCP (a standard protocol for connecting AI models to outside tools and data).
The result is that every agent knows the current build state — what files have changed, what decisions were made — without being told repeatedly. That shared awareness cuts redundant LLM calls and wasted tokens. The pattern was posted with working examples, making it a practical starting point for teams or solo builders who want multiple agents to collaborate on a complex codebase instead of colliding.
Key points
- Agents share a common context so they don't duplicate work or contradict each other
- Hooks trigger automatic state updates at key moments in the workflow
- Skills package reusable logic once so every agent can call it without re-implementing it
- MCP gives all agents consistent access to the same external tools and data
- Fewer redundant LLM calls means lower token costs for multi-agent pipelines
Quick term guide
- AI agents
- AI agents are AI tools that can carry out steps toward a goal, not just answer once.
- reference
- Using a source to find information or confirm facts while working.
- multi-agent system
- A setup where several AI programs each take a specific role and work together to complete one larger task
- multi-agent
- A setup where several AI agents each handle a different subtask and work together to complete a larger goal.
- AI models
- The core brain or underlying program that powers an artificial intelligence tool.
- AI model
- A program that can understand prompts and produce text, code, or answers.
- token costs
- Token costs are the fees paid for the text an AI model reads and writes.
- token cost
- The money or usage spent when sending text to an AI model and getting text back.