This is a simple implementation of a token bucket rate limiter in Python.
This repository is the unedited workspace output of an autonomous multi-agent
run coordinated by corralai — a herd of local LLM
agents (this run: ollama:qwen2.5-coder:7b across 11 agents) that researched,
designed, built, tested, security-reviewed, and documented the directive:
Build a Python package 'ratelimit' implementing a token-bucket rate limiter: a RateLimiter class with allow(key) and configurable capacity and refill rate, plus docstrings, a usage README, and unittest tests; make 'python3 -m unittest' pass
The mission converged in ~223 seconds (11/11 tasks, 1 finding filed and addressed). No human wrote or edited the code in this repo — small-model rough edges and all; it is published as-is for transparency.
Watch the full recorded replay of this run (and others) at https://corralai.dev/recordings.