Hiding a multi-agent system behind one MCP tool cuts token bloat and drift
anilkr84 · reddit · 2026-07-26
The author proposes hiding a multi-agent system behind a single MCP tool call, arguing that orchestration should happen server-side rather than in the client context.
They outline three generations of MCP server design:
- Box of tools: the client plans and threads state across thin stateless functions
- Tools + skillpack: the client still does all the work, only with better instructions
- Orchestration behind the boundary: one thick tool, such as askagent(goal), delegates to a server-side multi-agent system
The claimed benefits are flatter token costs, less goal drift, smaller client schemas, client portability, and a frozen contract that survives sub-agent swaps. The trade-offs are higher latency per call, less visibility into the chain, and the need to operate a real agent system in production.
The post ends by asking whether others have shipped “agent behind a tool” setups and where opacity became a problem.
Related event: Architecture Insight: Hiding Multi-Agent Orchestration in a Single MCP Tool(2 posts)→
More from coding & agent
- AWS on agent prompt caching: explicit + implicit patterns to slash token costs — AWS ML Blog · 2026-10-27
- He gave his AI agent a phone number and email — then hired it as social media manager — heyshrutimishra · 2026-10-27
- Firecrawl launches Alexandria company-data API and announces $75M Series B — devdigest · 2026-10-27
- 'Let an agent run wild on vague goals and it makes your code worse 9 out of 10 times' — eptwts · 2026-10-27
- LangChain to demo the in-house design agent its team uses daily, this Wednesday — LangChain · 2026-10-27
- Biotech Agent Series Day 8: A Notebook Review Agent to Auto-Audit Lab Records — nlarusstone · 2026-10-27