Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Ralph Loop

ASecurity

Run a ticket queue across agent sessions with a self-verifying loop: one item per iteration, verification gates outside the worker, commits only from a verified tree on a topic branch.

46 stars
0 votes
0 copies
0 views
Added 10/1/2026
ai-agentsgogit

Works with

cli

Security Analysis

A100/100

Scanned 10/1/2026

$npx -y skills add agentskillexchange/skills --skill ralph-loop --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Ralph Loop?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Ralph Loop
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/agentskillexchange-ralph-loop/badge)](https://www.skillsdirectory.com/skills/agentskillexchange-ralph-loop)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
Files
SKILL.md
---
name: "Ralph Loop"
slug: "ralph-loop"
description: "Run a ticket queue across agent sessions with a self-verifying loop: one item per iteration, verification gates outside the worker, commits only from a verified tree on a topic branch."
verification: "listed"
source: "https://github.com/Thingscorp/skills/tree/main/skills/ralph-loop"
category: "Templates & Workflows"
framework: "Custom Agents"
---

# Ralph Loop

Run a queue of vertical tickets across agent sessions with a self-verifying loop. Live agent context is disposable; durable state lives in `.ralph/` (plan, items, prompt, loop state, append-only progress log). One iteration completes exactly one item — no batching, no drive-bys.

The core discipline: never let the worker verify its own work. Each iteration spawns one fresh worker for a single item under a strict worker contract (change only that item's allowed paths, run the gate, never set `passes: true`, never touch `.ralph/loop.md`). When the worker returns, the engine saves the candidate as an immutable Git tree and runs the saved baseline gate plus the current gate **outside** the worker, in a detached worktree of that exact tree. Acceptance requires exit 0 and a final line of `ALL GATES PASS`. On pass, the engine commits hook-free (Git plumbing, not hook-rewritable `git commit`) from the verified tree, marks the item passed, appends an engine progress entry, and releases the `.git/ralph.lock` lock. On fail, it writes a recovery ref, restores the last verified branch state, and records the stop reason. A `runtime_contract` declares the verification gates, topic branch, protected paths, and the non-negotiable rules: one item per iteration, commit required, progress appended, never push without owner approval. Optional judgment gates (select, verify, triage, blocker, continue) let a small model answer the semantic questions so the frontier model doesn't have to; gate errors fall back to FIFO with the fallback stated. This is the portable alternative to runner CLIs: a ticket-queue contract any harness can follow, proven running real maintenance queues.

## Installation

No source-backed install or usage instructions could be extracted automatically. Review the upstream project before running this skill in a sensitive workflow.

- Source: https://github.com/Thingscorp/skills/tree/main/skills/ralph-loop

## Source

Upstream skill: https://github.com/Thingscorp/skills/tree/main/skills/ralph-loop — part of the MIT-licensed Thingscorp/skills library (15 operational SKILL.md playbooks). Companion skills: ralph-plan (verified plans + items, the upstream of this loop), ralph-swarm (parallel lanes), to-spec, quality-loop.

Attribution

agentskillexchangeagentskillexchange
View sourceSee grades on GitHubMore from agentskillexchange →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

698461 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →