Brian Hung
Harness
Recursive language models and a continual harness, as MCP servers any agent can use.
rlm_deep_querythe context never enters the model windowquestionanswerrlm_store_contextfull textcodestdout, first 4,000 charssliceanswerMCP hostClaude Code, Codexcontext storerlm_store_contextroot modelgenerateText tool loopPython kernelPROMPT holds the contextsub-callone depth downThe answer is the model's reply, or FINAL from the kernel when it is too long to write out.Below the depth limit a sub-call gets its own kernel; at the limit it is one plain model call.On Docker and Cloudflare, the model calls a recurse tool instead.