claudeloop¶
Onion-architected, autonomous Claude Code session runner and full Anthropic SDK CLI — never blocks on a human, distinguishes an exhausted rate-limit window from exhausted credits, and resumes safely across usage windows.
What problem this solves¶
Claude Code sessions hit usage limits. A claude -p invocation ending
doesn't tell you whether the task finished or just that turn did. And
when a rate limit rejects you, you can't tell from the outside whether
waiting will ever help — a five-hour window will reset on its own; an
exhausted credits balance never will, no matter how long you wait.
claudeloop exists to get all three of those distinctions right,
automatically, so you can hand it a plan and walk away — including handling
the case where you top up your account's credits while it's mid-wait.
Where to go next¶
| I want to... | Read |
|---|---|
| Install it and run my first autonomous session | Getting started |
| Understand how rate limits and credits are told apart | Rate limits vs. credits |
| Understand why it can never get stuck waiting on me | Never blocking |
| See the full system design | Architecture overview |
| See why a specific hard call was made | Architecture decision records |
| Contribute code | Development guide |
| See the original plans this project was built from | Plans |
Project status¶
Pre-1.0. Milestone M1 (the pure domain core) is complete — a fully
tested set of value objects and a pure state machine that model every hard
decision the runner makes. Milestones M2–M5 (the actual agent
integration, the CLI, the generated REST surface, and final polish) are
roadmap. See plans/architecture-and-roadmap.md
for the build order.
License¶
MIT. See LICENSE.