actionable ai

Free guide

Let Claude work for 8 hours while you sleep. One command. Zero babysitting.

The exact condition-builder prompt I use, 5 ready-to-paste long-running goal templates, the drift checkpoint pattern that stops 90% of long-run failures, and the Codex parity context most coverage skips.

  • The /goal-condition-builder prompt (Anthropic's 3-part framework: measurable end state + verifiable check + drift constraints), works identically for Claude /goal and Codex /goal
  • 5 ready-to-paste goal templates (test migration, issue backlog drain, file splitting, doc backfill, dependency upgrade)
  • How Claude /goal compares to Codex's /goal, where each pulls ahead, where they're at parity, why the condition is the entire game
  • Real receipts from the 8-hour run in the video: git log, test output, PR view

Instant access

Get the free guide

The exact prompt and 5 templates for running your first autonomous Claude Code /goal tonight, plus the full Codex parity breakdown. Anthropic's own framework, the drift fix that keeps long runs from wandering, and the competitive context nobody else is putting in one place.

One-time signup. Every guide in the vault unlocks with it.

A look inside

What's inside

01

What You'll Learn

The exact condition builder prompt that makes /goal runs actually finish 5 ready to paste long running goal templates (test migration, issue backlog, file splits, doc backfill, dep upgrades) The drift checkpoint pattern that stops 90% of long run failures How Claude /goal compare

02

What /goal Actually Does

You type one command, write a completion condition, and Claude keeps working across turns until the condition is met.

03

How Claude /goal Compares To Codex's /goal

OpenAI shipped /goal in Codex on May 3, 2026, about ten days before Anthropic shipped it in Claude Code.

The rest of the guide

  1. The /goal Condition Builder PromptMost people write conditions Claude cannot verify, and the run bails early or drifts.
  2. The Drift Checkpoint PatternThe number one failure mode for long /goals is the agent solving a different problem than the one you set.
  3. 5 Ready To Paste Goal TemplatesEach template is a complete /goal condition you can adapt by swapping the bracketed values.
  4. Scoping For 6 To 12 Hour FinishesTwo questions to ask before setting a goal.
  5. The 4 Most Common Condition MistakesVague success: "the code is cleaner." Rewrite to "npm run lint exits 0 and no file exceeds 400 lines" Missing constraints: "all tests pass." Add "and no implementation file outside src/auth is modified" No turn or time bound: the condition can technically be satisfied forever.
  6. Real Receipts From The 8 Hour RunA real /goal run finished a test migration that had been sitting in the backlog.
  7. Your First /goal Tonight (30 Minute Plan)Update Claude Code to 2.1.139 or later.
  8. Further ReadingAnthropic /goal docs: code.claude.com/docs/en/goal OpenAI Codex 25 hour run blog: developers.openai.com/blog/run-long-horizon-tasks-with-codex Cursor on scaling agents: cursor.com/blog/scaling-agents Anthropic long running Claude (scientific computing): anthropic.com/research/lon
  9. Go DeeperPrefer a printable version?

About the author

Cooper Simson

Founder of Actionable AI. 30,000+ operators read the newsletter.

Good to know

Questions, answered

Is this guide free?

Yes. Enter your details and you can open the full guide immediately.

Do I need to know how to code?

No. The guide explains the work in practical steps and calls out technical details only when they matter.

What happens after I enter my email?

The guide opens from this page. If the guide email is enabled, you will also receive a link for later.

Related

View all guides

guide

The 1M Context Playbook: Get Clean Structured Answers Out Of 500K Plus Tokens

The prompt template and workflows that turn Claude's new 1M token window from a headline into a real unlock.

Get the guide

guide

The Workflow File: How To Run 100 Claude Agents In Parallel (Anthropic's Unannounced Update)

The JavaScript file plus 3 ready-to-paste templates for Anthropic's unannounced Claude Code /workflows feature. Sentry triage, dead-code sweep, personalized outreach. Plus the env variable to turn it on.

Get the guide

guide

The Loop Starter Kit (3 AI Loops + Commands)

The 3 AI loops to set up first (inbox, overnight builder, viral content), the exact Claude Code commands to paste in, and a link to the full 100+ loop catalog.

Get the guide

guide

The From-Scratch /route Setup Guide - Fable 5 + GPT-5.6 Sol

The from-scratch guide to wiring Fable 5 and GPT-5.6 Sol into one plan-build-review loop.

Get the guide

guide

Run Kronos: Full Walkthrough

Step-by-step walkthrough for installing and running the Kronos Claude skill. Every command and prompt is copy-paste ready.

Get the guide

guide

The Claude Code /checkup Guide

The free one-command tune-up for Claude Code: the exact /checkup command, how to run it safely, what all 7 cleanups do, and when to use it.

Get the guide