2026-10-11 18:03 UTC

Routed’s maintainer claims its released local hybrid router selects AI-agent skills in under 20 milliseconds without consuming model tokens, potentially removing model-call cost and latency from skill routing.

state: expiredheat: lowuncertainty: highknownscott: lowagent-harnesses skill-routing local-inferencebshea-1

What is this?

Routed is presented in the supplied case as a released local hybrid router for selecting AI-agent skills, attributed to maintainer bshea-1. The maintainer’s HN submission reportedly claims routing in under 20 milliseconds without consuming model tokens, potentially avoiding a model call for skill selection. None of the supplied search snippets directly documents Routed or independently verifies its release, implementation, or performance; they discuss other routing systems, and do not establish that Routed routes between local and cloud models.

Why it matters to Scott

Routed’s claimed model-call-free skill selection illustrates Scott’s existing Placement Judgment position and Deterministic agent control plane design, rather than establishing a new architectural claim. The supplied material offers only maintainer testimony—not implementation details, routing-quality evidence, or verified latency—so it does not yet justify changing his routing approach; related radar pages track discovery costs, but not Routed itself.
ip:concept.placement-judgmentdev:concept.deterministic-agent-control-planeradar:concept.agent-skillsradar:mcptoon-tool-discovery-compression
queries asked of Scott's wikis
  • agent harness skill discovery and routing
  • deterministic dispatch versus LLM skill selection
  • routing latency and model-call token costs
  • local hybrid retrieval for tool selection
  • skill routing accuracy and fallback evaluation

Measured heat

no measured readings yet — the hourly heat pass fills this in

How the heat travelled

no chain yet — the hourly chain pass fills this in

Evidence (7) — ⭐ canonical anchor

sourceobjectauthorscorecomments
🟧 hnRouted: Local, zero-token hybrid router for AI agent skills (<20ms)bshea120
🟧 echo.github ⭐The maintainer’s HN submission describes Routed as a local, zero-token hybrid router for AI agent skills with latency below 20 milliseconds.bshea-1——
🟠 redditI built a local-first hybrid router for AI Agent Skills (sub-20ms, zero tokens, runs on CPU)
ClaudeAI
iAmQubick31
🟠 redditI built a local-first hybrid router for AI Agent Skills (sub-20ms, zero tokens, runs on CPU)
ClaudeAI
iAmQubick01
🟠 redditI built a local-first hybrid router for AI Agent Skills (sub-20ms, zero tokens, runs on CPU) [P]
MachineLearning
iAmQubick01
🟠 redditI built a local-first hybrid router for AI Agent Skills (sub-20ms, zero tokens, runs on CPU)
ClaudeAI
iAmQubick12
🟠 redditI put together a live demo for my local-first hybrid AI skill router (Offline, zero tokens)
ClaudeAI
iAmQubick11

Interpretation history

Decision trace