Umputun Ralphex
- 1.4k repo stars
- Updated July 26, 2026
- umputun/ralphex
umputun-ralphex is a Claude Code plugin that runs an extended Ralph loop for autonomous plan creation, execution, and monitoring.
About
umputun-ralphex is a Claude Code plugin that implements an extended Ralph loop for autonomous, plan-driven work. Catalog description and keywords highlight creation, execution, and monitoring— the pattern developers use when they have a written plan and want the agent to keep working through it with less manual babysitting than chat-by-chat prompting. It fits multi-step product builds, migration projects, and operational fix cycles where progress tracking matters as much as code output. You should still define guardrails, tests, and rollback paths; autonomous loops amplify both speed and risk on a one-person repo. Think of ralphex as execution infrastructure for AI-native PM: not a brainstorming skill, not a marketplace of snippets, but a monitored loop that extends classic Ralph automation for Claude Code sessions that run across build and into operate-style monitoring.
- Extended Ralph loop for autonomous AI-driven plan execution inside Claude Code
- Emphasizes plan creation, hands-off execution passes, and ongoing monitoring
- Single-plugin listing (pluginCount 1) from umputun/ralphex (~925 catalog stars)
- Suited to long-horizon solo projects where the agent iterates until the plan converges
Umputun Ralphex by the numbers
- Data as of Jul 28, 2026 (Skillselion catalog sync)
/plugin install umputun-ralphex@umputun/ralphexAdd your badge
Show developers this plugin is listed on Skillselion. Paste this into your README.
| repo stars | ★ 1.4k |
|---|---|
| Last updated | July 26, 2026 |
| Repository | umputun/ralphex ↗ |
What it does
Install ralphex when you want an extended Ralph loop in Claude Code for autonomous plan creation, execution, and monitoring on AI-driven builds.
Who is it for?
Experienced developers comfortable supervising autonomous agent loops on well-defined plans and repos.
Skip if: Beginners who need step-by-step human approval on every file change or teams without test and rollback discipline.
What you get
After you add ralphex, your agent can iterate on a plan with Ralph-style autonomy while creation and monitoring stay wired into the same Claude Code workflow.
- Ralph-style autonomous execution cycles tied to your plan
- Ongoing monitoring hooks between agent iterations
- Reduced manual re-prompting across long build or launch tasks
By the numbers
- Catalog lists pluginCount 1 for umputun/ralphex
- Catalog description: extended Ralph loop for autonomous AI-driven plan execution
- Catalog metadata reports approximately 925 GitHub stars (starsFetchedAt 2026-04-06)
Recommended Plugins
0pg Cc MarketplacePersonal Claude Code plugins collection
1shot Labs Squadron Comms PluginVoice communication system with ElevenLabs TTS for multi-agent coordination in Claude Code3
23blocks Os Ai MaestroAI Agent Orchestrator with Skills System - Give AI Agents superpowers: memory search, code graph queries, agent-to-agent messaging.733
3commas Io Commas ClaudeThis is a repo with shared agents and skills used company-wide4
884js Agent Plugins1 plugin
A5c Ai BabysitterBabysitter enforces obedience to agentic workforces and enables them to manage extremely complex tasks and workflows through deterministic, hallucination-free self-orchestration1.6kHow it compares
Autonomous Ralph execution loop, not a lightweight single-shot skill or external CI-only runner.
FAQ
Who is Umputun Ralphex for?
It is for Claude Code users who already trust plan-driven agent work and want Ralph-style autonomy with monitoring for extended tasks.
When should I use Umputun Ralphex?
Use it when you have a clear plan or backlog and want autonomous creation and execution iterations rather than one-off answers during build or launch prep.
How do I add Umputun Ralphex to my agent?
Install the umputun/ralphex plugin in Claude Code, configure the loop per repository documentation, and start from an explicit plan document the agent can execute and monitor.