Claude Code Autopause — Productivity skill for Claude Code
Claude Code skill that pauses your session at a usage-limit percentage you pick, checkpoints the task, and resumes by itself when the limit resets.
How to install Claude Code Autopause
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open mofchris/claude-code-autopause and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Claude Code Autopause does
Claude Code skill that pauses your session at a usage-limit percentage you pick, checkpoints the task, and resumes by itself when the limit resets.
Alternatives in Productivity
- Claudenightswatch — Autonomous task execution system for Claude CLI that monitors your usage windows and executes predefined tasks 352 ★
- Cezar — Cezar is a parallel coding agents orchestrator 174 ★
- Hybrid Auto — Generate PRD from task description and enter review mode 123 ★
README
claude-code-autopause
**Never lose work to the Claude usage limit again.**
A Claude Code skill that pauses your session at a percentage *you* pick, saves exactly where it was, and resumes by itself the moment the limit resets.
[](https://github.com/mofchris/claude-code-autopause/stargazers) [](LICENSE) [](#requirements) [](scripts/usage-guard.mjs)
The problem
You kick off a long task, walk away, and come back to a session that slammed into the 5-hour limit halfway through a refactor. Half the files are changed, tests never ran, and the model's working memory of what came next is gone. Or worse: you burned the whole weekly budget on Tuesday.
Usage monitors tell you the number. **This one acts on it.**
What it does
| Pauses at your number | 80% of the 5-hour window, 95% of the weekly window, whatever you choose. Per session if you want. |
| Blocks the expensive stuff only | Build tools (Bash, Edit, Write, subagents) are denied by a hook. Reading, searching and planning still work. |
| Saves the exact spot | A checkpoint file with the goal, what is done, the numbered next steps and how to verify. Nothing lives only in the model's head. |
| Resumes on its own | Schedules a one-shot wake-up for reset time, re-checks usage, reads the checkpoint, carries on. |
| Warns before it bites | As usage approaches the threshold the session is told to finish the current step cleanly, not sprint. |
| **Get |
Related Skills
Omarchy Auto Pilot4agents
Write a prompt now and send it later to a Claude Code, OpenCode, Codex, Gemini CLI, Cursor Agent or Pi session
Pantheon Deepwork
Start a heavy multi-phase task with persisted checkpoints, phased specialist dispatch, and Themis review gates
Summary
Produce a summary so another session can continue with zero context beyond CLAUDE.md, the codebase, and the su
Ve Vibe Board
Task board and session handoff for Claude Code agents. 14 MCP tools for projects, tasks, sessions, and activit
Claude Usage Bar
macOS menu bar widget that displays your Claude Code rate limit usage in real time
Next Decision
Drill the open decisions one at a time — most important first, four options, your pick, then wait. Usage: /nex
Related Agents
Session Handoff Writer
Writes and reads handoff documents so work continues across sessions or context resets, and keeps the cross-pr
Wave Babysitter
Babysits a wave of background agents: watches their output files, auto-resumes API/limit deaths ONCE per agent
Hard Reviewer
Independent review of a HARD change (one the main session made itself) by a different model from the one that