How do you route vague prompts to the right subagent? Devs trade production tips
CriticalJackfruit404 · reddit · 2026-09-30
A Reddit developer building a graph-based system with multiple subagents has a triage agent that routes requests (metrics questions to an analytics agent, raw event data to a raw-data agent). It works when users are explicit, but many real prompts are vague.
Two candidate fixes both fall short: stuffing markdown docs for every context into the triage agent risks bloat, token burn, and slower noisier routing as contexts grow, while just asking the user to clarify kills automation. They're asking for production-proven approaches to reliable multi-agent routing under ambiguous input.
More from coding & agent
- First DeepSeek Harness plugin ships: Qiaomu AI RSS brings AI news and podcasts into your coding agent — vista8 · 2026-09-30
- DoorDash launches corporate ordering MCP and CLI so AI agents can place orders, Vercel and Cognition early adopters — Scobleizer · 2026-09-30
- YC launches Video Arena, a free arena where AI models compete to code your video prompts — ycombinator · 2026-09-30
- openharness: One Tmux-Based TUI to Command Every Coding Agent on Every Machine — dee_hw · 2026-09-30
- Your MCP anonymizer is self-defeating if it takes code as an argument — Aggressive-Course-24 · 2026-09-30
- Mirage launches free Tesseract: agents design and animate editable visuals without Figma — chrisfirst · 2026-09-30