stateful-ralph
Skill3 starsChanged 18 days agoApache-2.0View on GitHub
Repeats the same task every round in one persistent worker session that remembers earlier rounds. Run with /baton:stateful-ralph "<task>".
In this repository17 other files it gives its agents.
- author-flowSkill
- cleanup-flame-chaseSkill
- cleanup-ralphSkill
- continue-loopSkill
- fixed-interruptSkill
- flame-chaseSkill
- flowsSkill
- gen-ideaSkill
- gen-planSkill
- goalSkill
- parallel-chase-prSkill
- parallel-chaseSkill
- protocolSkill
- ralph-loopSkill
- recursive-lean-proverSkill
- rlarSkill
- rlcrSkill
The file
Files
stateful-ralph
Arguments: $ARGUMENTS
Load the run protocol first (Skill tool → baton:protocol; if that fails, Read ${CLAUDE_PLUGIN_ROOT}/skills/protocol/SKILL.md) and follow it. This file only adds what is specific to stateful-ralph.
Idea
Like ralph-loop, but the worker keeps its session, so it remembers what it tried. It's faster on tasks where earlier reasoning matters, at the cost of the worker's context filling up and drifting.
Roles
| Role | subagent_type | Model | Session |
| worker | baton:worker | worker model | persistent, named <run-id>-worker |
Turn prompts
Round 1 (Agent call with name):
BATON stateful-ralph · turn 1/<max> · role worker · run <run-id>
TASK:
<task, verbatim>
You keep this session across rounds. Each round you will get the same TASK again. Make the most valuable progress you can this round and report.Round n > 1 (SendMessage to <run-id>-worker):
BATON stateful-ralph · turn <n>/<max> · role worker · run <run-id>
Round <n>. Same TASK:
<task, verbatim>
Continue from where you are. Check what is still missing before you add anything.Fallback: if the worker can't be reached, play the worker yourself in this session for the remaining rounds and say so in the log.
Stop
- The budget is used up, or a turn fails (protocol).
- Early stop, unless
--no-early-stop: thestreak:line showsdone x2or more andempty x1or more.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.