langchain - 💡(How to fix) Fix Feature: Support for non-probabilistic decision frameworks (hexagram state machines) in agent orchestration [1 comments, 2 participants]

Submission checklist

• This is a feature request, not a bug report or usage question.
• I added a clear and descriptive title that summarizes the feature request.
• I used the GitHub search to find a similar feature request and didn't find it.
• I checked the LangChain documentation and API reference to see if this feature already exists.
• This is not related to the langchain-community package.

Feature Description

I'd like to propose support for non-probabilistic decision frameworks in LangChain's agent orchestration — specifically, deterministic state machines inspired by I Ching hexagram mechanics.

Current approach: LangChain agents rely primarily on LLM-driven reasoning (ReAct, Plan-and-Execute, etc.) for decision-making.

Proposed extension: Allow pluggable decision backends that can combine LLM reasoning with deterministic state machines. For example:

• Map system metrics to a 6-dimensional vector
• Convert to a hexagram (64 possible states)
• Use hexagram transitions to drive agent behavior deterministically
• Fall back to LLM only when the state machine encounters ambiguity

This would enable auditable, traceable decision paths — you can explain exactly why an agent took each step.

Use Case

I built TaijiOS (https://github.com/yangfei222666-9/TaijiOS/tree/five-engines) — a working implementation with five engines:

1. Situation Engine — 6D vector analysis + deadlock breaking
2. Recovery Engine — 6-stage fault recovery state machine
3. Swarm Engine — Multi-agent coordination with commander selection
4. Learning Engine — Experience loop with weight decay
5. Persona Layer — Yin-yang balanced team composition

Validated with 80-event end-to-end demo, zero errors.

Alternatives Considered

1. Custom AgentExecutor subclass — works but doesn't integrate with LangChain's ecosystem
2. LangGraph state machines — closest existing solution, but designed for LLM-driven transitions
3. Pure LLM reasoning — loses determinism and auditability

Additional Context

Website: https://taijios.xyz Exam score: 83.6% on CozeAgent benchmark (top 21%)

Happy to contribute a PR if there's interest.