Concordat: Typed Sessions for LLM Agents
What it is
Concordat is a Python library that enforces finite-state protocols on LLM agents at runtime to ensure safety and correctness. It restricts available tools based on the current protocol state, rejects illegal calls before side effects occur, and manages token budgets across handoffs.
The problem it solves
LLM agents often ignore safety rules stated in prompts (e.g., skipping required confirmations) because compliance relies on the model's probabilistic choice rather than hard constraints.
What you could use it for
- Enforcing mandatory customer confirmation steps in airline booking agents
- Preventing unauthorized account changes or refunds in financial services bots
- Implementing type-safe state machines for autonomous agent workflows
Why it matters
Significantly reduces the risk of consequential errors (like unauthorized refunds) in LLM agents by making safety rules structurally enforced rather than prompt-dependent, potentially enabling higher trust in autonomous systems.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.