Vibe Coding with AI Agents: method, prompts, skills and checklists
by PaperChaseWebb
A Markdown playbook for building real software with Claude Code, Codex or Claude: a seven-step method, project file templates, 18 prompts, five agent skills, release and hardware checklists, cost notes and 24 project ideas.
- Price
- $39
- Difficulty
- Intermediate
- Setup
- 20 min
- Format
- .md file
vibe-coding-with-ai-agents.md · 31 KB · instant download after checkout
Want this built for you instead? We can build it for you.
What you end up with
Your agent proves each stage works with tests, screenshots and measurements, and a fresh reviewer checks it before anything leaves your machine.
What you get
- 9 sections in one .md file, about 20 pages
- Seven-step method: spike, brief, stage, prove, review, gate, hand off
- BRIEF, AGENTS, DECISIONS, BLOCKERS and HANDOFF templates
- 18 prompts from first spike to release gate
- Five SKILL.md files: verifier, fresh critic, cross-model reviewer, handoff writer, release gate
- Checklists: before you push, before you release, reviving old hardware, guardrails
- Cost and context notes with a method to measure your own numbers
- Concept library of 24 projects with risks and first prompts
- Linked credits for all 33 sources we studied
Who it is for
People building apps, tools and drivers with an AI coding agent who want it to check its own work and stop before anything irreversible.
Works with
Solid chips: the agent is named in this listing's own tools list. Dashed chip: Plain Markdown, so it also pastes into other agents such as Grok Bot, Hermes, OpenClaw or Muse. We list those as compatible by format only, not as tested.
Tools needed
Cost to run: Free. Uses the agent plan you already pay for.
Time to result: Project files and first stage in one evening
Sample prompts, free to use
For issue #[n]: write a test that reproduces the bug and show me it failing. Then make the smallest fix, show the same test passing, and run the full suite.
A fix without a failing test first is a guess.
This chat is getting long. Write HANDOFF.md: goal right now, what was done, open tasks in order, files changed, traps. Update DECISIONS.md and BLOCKERS.md. Keep it under one screen.
A fresh session with the reasoning beats a slow one with everything.
Free preview of the file
# Vibe Coding with AI Agents A PaperChaseWebb pack, version 1.0. Written for Claude Code, Codex and Claude, with notes for Cursor where a source used it. Plain Markdown, so any other agent can read it, but those are the agents the prompts and skills were worded for. This pack does not write your app. It is the working method we pulled out of 33 public build reports from October 2026: people who got real software running with an agent doing most of the typing, and the habits that kept it from falling apart. Apps, firmware, drivers, browser extensions, a 500-piece gallery, a pool of PR workers. Every source is linked and credited in `08-sources.md`, and everything here is in our own words. ## Files | File | What it is for | |---|---| | 01-method.md | The seven-step loop: spike, brief, stage, prove, review, gate, hand off | | 02-project-files.md | Fill-in BRIEF.md, AGENTS.md, DECISIONS.md, BLOCKERS.md and HANDOFF.md templates | | 03-prompts.md | 18 prompts from first spike to release gate | | 04-agent-skills.md | Five drop-in SKILL.md files: verifier, fresh critic, cross-model reviewer, handoff writer, release gate | | 05-checklists.md | Before you push, before you release, reviving old hardware, agent guardrails | | 06-cost-and-context.md | How to measure what a task costs and keep long sessions fast | | 07-concept-library.md | 24 project ideas with the build pattern, the main risk and a first prompt | | 08-sources.md | Every source we used, linked and credited |
Worth knowing
- The pack is a method and prompt library. It does not include an app, code or hosting.
- Tools and model limits change. Measure your own numbers.
- Never let an agent deploy, submit, merge or spend without asking you.
Questions
Which agents is it written for?
Claude Code, Codex and Claude. The skills use their file layouts. Other agents can read the Markdown, but we have not written for them.
Is there a free sample?
Yes. The decision log, blocker log and handoff skill are free on their own listing.
Is any of it copied?
No. We read 33 public build reports and wrote everything in our own words. Sources are linked and credited inside.
Details
Version 1.0 · Updated Oct 25, 2026 · Markdown version of this page
License: Personal and commercial use. Your apps are yours. Do not resell or republish this pack. Download terms
Related
Hand off agent sessions with decision logs (free sample)
Two small files and one skill that let a fresh Claude Code or Codex session pick up the reasoning behind your code, not just the code.
Debug a vibe-coded app without breaking it further
12 prompts that make your AI coding tool find the real cause of a bug, change one thing at a time and prove the fix before it moves on.
AGENTS.md, CLAUDE.md and Cursor rules starter files
Six fill-in instruction files for coding agents (base, Next.js plus Supabase, FastAPI, browser game, CLAUDE.md pointer, scoped Cursor rules) and 8 prompts to write, trim and test your own.