Production-grade prompts and workflows
for the AI coding tools you actually use.

34 hand-written system prompts, Cursor rules, agent configs, and workflows for senior developers using Claude Code, Cursor, Copilot CLI, Aider, and Continue. Plus a 30-page operator's guide. No fluff. No AI slop. Built by an engineer who actually ships.

12 system prompts 8 Cursor rule sets 8 agent configs 6 workflows 30-page guide

$29 one-time. Lifetime updates. Money-back if it isn't worth it.

Why this exists

AI coding tools are now part of professional software development. Most users get maybe 30% of the available value because they treat the tool as a fancy autocomplete. The other 70% comes from how you frame tasks, what context you load, what guardrails you set, and what workflows you run.

You can either spend six months figuring it out on your own — burning tokens, shipping bugs, undoing agent mistakes — or you can start where someone else's six months ended.

This pack is the kit a senior engineer would assemble for themselves, if they had the time. Hand-written. Opinionated. Battle-tested. No "comprehensive guide to prompt engineering" theory.

What's inside

34 assets across four categories, plus a 30-page operator's guide. Every asset is hand-written, opinionated, and immediately usable.

01 — System Prompts

12 production-grade roles

Drop-in system prompts that give your agent a consistent, professional persona for the work you're doing.

  • The Senior Engineer
  • The Code Reviewer
  • The Debugger
  • The Refactorer
  • The Architect
  • The Test Writer
  • The Performance Engineer
  • The Security Reviewer
  • The Documentation Writer
  • The API Designer
  • The Migration Engineer
  • The Incident Commander
02 — Cursor Rules

8 .cursorrules for major stacks

Ready-to-drop .cursorrules files that encode the conventions Cursor's agent should follow for each stack.

  • Next.js + TypeScript (App Router)
  • Python + FastAPI
  • Go (idiomatic stdlib-first)
  • Rust
  • React + Vite (SPA)
  • Python data science / ML
  • Terraform / IaC
  • SQL (PostgreSQL-first)
03 — Agent Configs

8 CLAUDE.md / AGENTS.md files

Repo-level configurations that calibrate Claude Code, Copilot CLI, and Cursor agent mode to your project's context.

  • Production codebase
  • Solo indie / small project
  • Open source library maintainer
  • Enterprise / regulated codebase
  • Greenfield (day zero)
  • Legacy codebase
  • Data / ML repository
  • Long-running autonomous agent
04 — Workflows

6 multi-step playbooks

Step-by-step workflows for common dev tasks. Each one tells you what to ask, what to verify, and what to never let the agent do.

  • Thorough PR Review
  • The Debug Loop
  • Spec → Pull Request
  • Surgical Refactor Loop
  • Bootstrap a Test Suite for Untested Code
  • Codebase Onboarding

+ The 30-page Operator's Guide

~9,000 words on getting professional results from AI coding tools.

Nine parts covering: the mental model, how to prompt without writing essays, the 12 most common mistakes (and their cures), workflow patterns that work, when NOT to use the agent, setting up a project for AI agents, anti-patterns to retire, calibrated expectations, and the ten habits of high-leverage agent users.

Free preview

11 of the 34 assets are publicly readable so you can judge the quality before paying. If these don't read like "I'd actually use this," the rest won't either.

Who this is for

Built for:

  • Senior / staff engineers who already pay for AI coding tools
  • Tech leads codifying conventions for their team
  • Solo devs and indie hackers shipping with AI agents daily
  • Developers tired of "10x prompts" listicles that are 0.1x prompts
  • People who want their agents to behave consistently, not creatively

NOT for:

  • People who've never used Claude Code, Cursor, or similar
  • People who want generic "100 ChatGPT prompts" of any kind
  • Beginners looking for a tutorial on programming
  • People who think AI will replace senior engineers next quarter
  • People who think AI is a fad

Get the full pack

34 assets + 30-page operator's guide. One-time $29. Lifetime updates. Try it for 14 days — if it doesn't earn its cost, full refund, no questions.

Get the Playbook — $29 $29  ·  instant download  ·  lifetime updates

About the author

This pack was built by an engineer who's been writing code professionally for over a decade and using AI coding tools in production for the last two years. The prompts, rules, and workflows here are not theoretical — they're the same setups used to ship real software.

No part of this pack is generic AI-generated filler. Every asset was written by hand, refined through real use, and pruned ruthlessly. Opinion-rich and length-pruned.

— Josh Alexy

FAQ

What format is the pack delivered in?

Markdown files. Every prompt, rule, config, and workflow is a .md file you can paste, drop into a repo as CLAUDE.md / .cursorrules, or read as documentation. The operator's guide is included as both markdown and PDF.

Will this work with [my specific AI tool]?

The system prompts work with any model that accepts a system prompt (Claude, GPT, Gemini, Llama, etc.). The Cursor rules work in Cursor as .cursorrules. The agent configs work with Claude Code (CLAUDE.md), Copilot CLI (AGENTS.md), Cursor agent mode, and Aider/Continue with minor adaptation.

Why $29 and not free / not $99?

$29 means people who buy it will actually use it. Free packs get hoarded. $99 packs assume hours of bespoke consulting. This is priced as "less than one hour of senior eng time" — which is roughly the breakeven point if the pack saves you one prompt-engineering rabbit hole.

How is this different from cursor.directory or other public prompt collections?

Three differences: (1) Curation — every asset is hand-written and opinionated rather than community-submitted boilerplate. (2) Depth — these are real working prompts with specific rules, not one-line "act as a senior engineer" snippets. (3) Coherence — the assets are designed to compose; the system prompts and workflows reference each other.

What if I don't like it?

14-day money-back guarantee, no questions asked. Email and you get refunded.

Do you update the pack?

Yes. When new AI coding tools become popular or major versions change conventions, the pack gets updated. Lifetime updates are included in the purchase.

Can I share this with my team?

The license allows internal team use — drop it in a private team repo, share with coworkers. Public re-distribution (e.g., reposting on a blog, selling as your own) is not allowed.

What if my company will pay for it?

Email me (link in footer) and I'll send an invoice with a business-style receipt. Companies usually expense this without trouble — it's significantly cheaper than an hour of a senior engineer's time.