

Also from Kynth Studios


Also from Kynth Studios


Also from Kynth Studios
123456# Ralph Wiggum Plugin Help78Please explain the following to the user:910## What is the Ralph Wiggum Technique?1112The Ralph Wiggum technique is an iterative development methodology based on continuous AI loops, pioneered by Geoffrey Huntley.1314**Core concept:**15```bash16while :; do17 cat PROMPT.md | claude-code --continue18done19```2021The same prompt is fed to Claude repeatedly. The "self-referential" aspect comes from Claude seeing its own previous work in the files and git history, not from feeding output back as input.2223**Each iteration:**241. Claude receives the SAME prompt252. Works on the task, modifying files263. Tries to exit274. Stop hook intercepts and feeds the same prompt again285. Claude sees its previous work in the files296. Iteratively improves until completion3031The technique is described as "deterministically bad in an undeterministic world" - failures are predictable, enabling systematic improvement through prompt tuning.3233## Available Commands3435### /ralph-loop <PROMPT> [OPTIONS]3637Start a Ralph loop in your current session.3839**Usage:**40```41/ralph-loop "Refactor the cache layer" --max-iterations 2042/ralph-loop "Add tests" --completion-promise "TESTS COMPLETE"43```4445**Options:**46- `--max-iterations <n>` - Max iterations before auto-stop47- `--completion-promise <text>` - Promise phrase to signal completion4849**How it works:**501. Creates `.claude/.ralph-loop.local.md` state file512. You work on the task523. When you try to exit, stop hook intercepts534. Same prompt fed back545. You see your previous work556. Continues until promise detected or max iterations5657---5859### /cancel-ralph6061Cancel an active Ralph loop (removes the loop state file).6263**Usage:**64```65/cancel-ralph66```6768**How it works:**69- Checks for active loop state file70- Removes `.claude/.ralph-loop.local.md`71- Reports cancellation with iteration count7273---7475## Key Concepts7677### Completion Promises7879To signal completion, Claude must output a `<promise>` tag:8081```82<promise>TASK COMPLETE</promise>83```8485The stop hook looks for this specific tag. Without it (or `--max-iterations`), Ralph runs infinitely.8687### Self-Reference Mechanism8889The "loop" doesn't mean Claude talks to itself. It means:90- Same prompt repeated91- Claude's work persists in files92- Each iteration sees previous attempts93- Builds incrementally toward goal9495## Example9697### Interactive Bug Fix9899```100/ralph-loop "Fix the token refresh logic in auth.ts. Output <promise>FIXED</promise> when all tests pass." --completion-promise "FIXED" --max-iterations 10101```102103You'll see Ralph:104- Attempt fixes105- Run tests106- See failures107- Iterate on solution108- In your current session109110## When to Use Ralph111112**Good for:**113- Well-defined tasks with clear success criteria114- Tasks requiring iteration and refinement115- Iterative development with self-correction116- Greenfield projects117118**Not good for:**119- Tasks requiring human judgment or design decisions120- One-shot operations121- Tasks with unclear success criteria122- Debugging production issues (use targeted debugging instead)123124## Learn More125126- Original technique: https://ghuntley.com/ralph/127- Ralph Orchestrator: https://github.com/mikeyobrien/ralph-orchestrator128
One repository carrying more than one format is the comparison this product exists for: does anyone actually write different content in each file, or is one a copy of the other?
| Repository | Format | Stack | Covers | Score | Changed |
|---|---|---|---|---|---|
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-build.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-code-simplify.mdc · 51 | Cursor rules | testing-strategy | 30/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-plan.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-review.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-ship.mdc · 51 | Cursor rules | testing-strategygitdeploymentdo-not | 61/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-spec.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-cursor/.cursor/rules/cmd-test.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-forgecat/AGENTS.md · 51 | AGENTS.md | lint-formatstylearchdo-not | 73/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/addyosmani/agent-skills/for-forgecat/CLAUDE.md · 51 | CLAUDE.md | teststylearchagent-behaviour | 70/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-code/anthropics_claude-code_ralph-wiggum/for-cursor/.cursor/rules/cmd-cancel-ralph.mdc · 51 | Cursor rules | no sections | 16/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-code/anthropics_claude-code_ralph-wiggum/for-cursor/.cursor/rules/cmd-ralph-loop.mdc · 51 | Cursor rules | no sections | 22/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_agent-sdk-dev/for-cursor/.cursor/rules/cmd-new-sdk-app.mdc · 51 | Cursor rules | setupstylearchdocs | 76/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_claude-md-management/for-cursor/.cursor/rules/cmd-revise-claude-md.mdc · 51 | Cursor rules | agent-behaviour | 50/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_code-review/for-cursor/.cursor/rules/cmd-code-review.mdc · 51 | Cursor rules | testing-strategygit | 35/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_commit-commands/for-cursor/.cursor/rules/cmd-clean_gone.mdc · 51 | Cursor rules | no sections | 60/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_commit-commands/for-cursor/.cursor/rules/cmd-commit-push-pr.mdc · 51 | Cursor rules | stylegit | 44/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_commit-commands/for-cursor/.cursor/rules/cmd-commit.mdc · 51 | Cursor rules | style | 44/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_example-plugin/for-cursor/.cursor/rules/cmd-example-command.mdc · 51 | Cursor rules | lint-formatstyleagent-behaviour | 58/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_feature-dev/for-cursor/.cursor/rules/cmd-feature-dev.mdc · 51 | Cursor rules | stylearchgit | 56/100 | today | |
| nota-america/forgecat-agent-profilesprofiles/anthropics/claude-plugins-official/anthropics_claude-plugins-official_plugin-dev/for-cursor/.cursor/rules/cmd-create-plugin.mdc · 51 | Cursor rules | testlint-formatstylearch+3 | 61/100 | today |
Same format, overlapping stack, ranked by quality.
| Repository | Format | Stack | Covers | Score | Changed |
|---|---|---|---|---|---|
| hiromaily/go-crypto-wallet.cursor/rules/typescript.mdc · 126 | Cursor rules | setupbuildtestlint-format+6 | 100/100 | 14 days ago | |
| TechSquidTV/Hermes.cursor/rules/10-hermes-api.mdc · 46 | Cursor rules | testlint-formatstylearch+5 | 100/100 | 14 days ago | |
| dodgecfr/combatfilms-webapp.cursor/rules/frontend.mdc · 0 | Cursor rules | setuptestlint-formatstyle+7 | 99/100 | 14 days ago | |
| deifos/clipmira-subtitles.cursor/rules/frontend.mdc · 1 | Cursor rules | setuptestlint-formatstyle+7 | 99/100 | 14 days ago | |
| markstev/mark-starter.cursor/rules/frontend.mdc · 0 | Cursor rules | setuptestlint-formatstyle+6 | 99/100 | 14 days ago | |
| Allymahmoud/case-intake-platform.cursor/rules/frontend.mdc · 0 | Cursor rules | setuptestlint-formatstyle+7 | 99/100 | 14 days ago | |
| langflow-ai/langflow.cursor/rules/docs_development.mdc · 153k | Cursor rules | setupbuildtestlint-format+7 | 97/100 | 14 days ago | |
| bybren-llc/safe-agentic-workflow.cursor/rules/10-backend-python.mdc · 399 | Cursor rules | testlint-formatstylegit+4 | 97/100 | today |
A badge carrying the measured quality of the strongest agent config file in this repository, out of 100. It reads from this index every time somebody loads your page, so it changes when the measurement changes and there is nothing to keep up to date. Free, no account, and the value is not something you or we can set by hand.
[](https://rulestack.kynth.studio/configs/nota-america-forgecat-agent-profiles-profiles-anthropics-claude-code-anthropics-claude-code-ralph-wiggum-for-cursor-cursor-rules-cmd-help)Would rather not hotlink us? Every badge is also served in shields.io’s endpoint schema, so shields renders the image and your readers never talk to our domain:
Published by Toolproof, the masthead over this index and eight others. The method behind the number is at toolproof.kynth.studio/methodology, and the whole thing is readable as JSON with no key at /api.