Overview & result
Lahfir reports using a deterministic Claude Code hook to ask Jev which local coding harness or model tier should take a task before delegation.
From the original post
Jev solved local harness/model routing
I use a combination of Claude Code, Codex and Opencode as my local agentic stack and routing to other harnesses was always enforced in the system prompt/rules
With a deterministic hook that Claude Code can decide before delegation, Jev helps to route to the right harness/model based on the task, and it's pretty accurate based on the intensity/intelligence of the task
> Mechanical tasks get routed to Haiku
> Intelligent ones to Opus sub-agents
> Long-running implementation work to external harness
How Jev fits in the loop
- Claude Code encounters a task before delegation.
- Jev assesses task intensity and selects a bounded routing destination.
- The hook delegates mechanical work to Haiku, higher-intelligence work to Opus sub-agents and long-running implementation to an external harness, as described by the author.
How to reproduce
- Read the original demo and identify the pre-delegation hook and three routing destinations.
- No public repository or install instructions were linked in the source post; implementation cannot yet be reproduced from this post alone.
Why this build matters
Shows an agent-harness routing boundary, while leaving implementation details and performance unverified.
Limitations
- The source provides a demonstration and routing outline, not public code or independent measurements.