ralph-loop

An agent skill by cursor, from cursor/plugins. Tags: ai, automation, development, multi-agent, workflow.

What it does

Start a Ralph Loop for iterative self-referential development. Use when the user asks to run a ralph loop, start an iterative loop, or wants repeated autonomous iteration on a task until completion.

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add cursor/plugins --skill ralph-loop

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/cursor/plugins
cp -r plugins/ralph-loop/skills/ralph-loop ~/.claude/skills/ralph-loop

Safety box score

Grade C, moderate risk: 14 of 100 risk points, rated 2026-10-16. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill sets up a self-feeding iterative loop that keeps reusing the same prompt across turns and writes loop state into `.cursor/ralph/scratchpad.md`. The main risk is autonomy: with the default `max_iterations: 0`, it can continue indefinitely and consume time/resources while modifying project-local state.

Category levels, from 0 (none) to 3 (high)
Category	Level	Why
Code & shell execution	0 (none)	No shell commands, scripts, or code execution are shipped; the file only describes a loop workflow in markdown.
Network access	0 (none)	The skill does not download from or send data to any remote host.
Filesystem reach	1 (low)	It creates `.cursor/ralph/` and writes `.cursor/ralph/scratchpad.md`, but the write is scoped to the working tree rather than system paths.
Secrets & credentials	0 (none)	It does not request API keys, passwords, tokens, or other secret material.
Privilege & persistence	0 (none)	No sudo, global installs, services, shell profile edits, or OS-level changes are requested.
Instruction hijack surface	0 (none)	There is no hidden text, external prompt fetching, or instruction to ignore higher-priority rules; it only defines a local workflow.
Irreversible actions	3 (high)	The skill is explicitly for an iterative self-referential loop; it says the stop hook re-feeds the same prompt each turn and the default `max_iterations` is `0` (unlimited).
Transparency	0 (none)	The repository contains a single readable markdown file and no binaries, generated blobs, or unreviewable installers.

Before installing

Decide whether you want an agent that can recurse across turns; if not, do not use this skill.
Set a finite `max_iterations` before enabling it.
Check that writing `.cursor/ralph/scratchpad.md` in the repo is acceptable for your workflow.

Rated 2026-10-16 by an automated review of 1 file from the skill's repository.

Source

Repository: cursor/plugins (all skills from this repository)
Path: ralph-loop/skills/ralph-loop/SKILL.md
Branch: main
Collection: cursor
Updated: 2026-09-23

orchestrate — Use only when the user explicitly types `/orchestrate <goal>` to decompose a large task.
continual-learning — Orchestrate continual learning by delegating transcript mining and AGENTS.md updates to `agents-memory-updater`.
swarm — Fan out N parallel workers, drain them, and return one report. Use for /swarm, 'swarm this', or parallel coverage, races, gauntlets, and exploration.
agents-build — Use to extend an existing agent project with memory, app integration, VPC, multi-agent, migration, model, browser, code interpreter, payments.
automate-me — Use for "automate me", "create/update/refresh my -mode skill", "turn/capture my preferences or working style into a skill".
cancel-ralph — Cancel an active Ralph Loop. Use when the user wants to stop, cancel, or abort a running ralph loop.