AaravKashyap12

AaravKashyap12 / advise-project-approach

Public

A Claude/Codex skill that researches the best way to build your project before you commit to the wrong stack.

agent-skills ai-agents claude-code claude-skills code-review
16
0
89% credibility
Found May 30, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

This project is a specialized helper for AI assistants that gives smart, evidence-based advice about how to build software projects. It acts like a wise friend who has seen many projects succeed and fail - before you commit to a tech stack or architecture, it researches comparable projects, evaluates your specific situation, and recommends the highest-leverage path forward. The skill adapts to three phases: pre-build (planning), mid-build (reviewing progress), and post-build (hardening before shipping). It emphasizes grounded recommendations over generic trends, requiring verifiable evidence for every claim it makes.

How It Works

1
💡 Discover the skill

Someone hears about this tool from a friend or online discussion - a helper that gives smart advice before you commit to building something the wrong way.

2
🔧 Add it to your AI assistant

You install the skill into your AI assistant using a simple one-line command, and it becomes available whenever you chat with it.

3
💬 Ask about your project idea

You describe what you want to build - maybe a bookmark manager, a web app, or an API - and your assistant springs into research mode.

4
Choose your phase
🏗️
Starting fresh

If you haven't built anything yet, it researches the best stack and comparable projects before you write a single line of code.

🔄
Already building

If you're halfway through, it reviews what you've done, spots real problems, and tells you what actually matters to fix next.

5
📊 Get evidence-based advice

The skill doesn't just guess - it checks real projects, verifies facts, and gives you recommendations backed by actual evidence instead of generic trends.

6
✅ Receive your build plan

You get a clear recommendation with comparable projects, architecture direction, and a prioritized plan - all organized and easy to understand.

🎯 Build with confidence

You now have a clear path forward, real examples to learn from, and the confidence that your approach fits your actual goals and constraints.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is advise-project-approach?

A Claude and Codex skill that acts like a senior engineer who has already made the mistakes you are about to make. Drop it into your AI coding assistant and it researches comparable projects, evaluates your proposed stack, and tells you whether you are about to build the wrong thing before you have written a line of code. The skill works across three phases: it advises before you start, audits mid-development, and reviews after completion. Built in Python with a YAML-based agent configuration, it installs via a simple npx command and requires no custom tooling.

Why is it gaining traction?

The hook is specificity over generic advice. Instead of telling you to "use Postgres" or "add Docker," it finds real projects solving similar problems, verifies their maintenance status, and explains exactly why a recommendation fits your constraints. It refuses to make claims without primary source evidence, which means no invented star counts or benchmark numbers. The skill also explicitly handles the three failure modes that plague AI code advice: recommending changes you already made, suggesting trends over fits, and confusing templates for production-ready solutions.

Who should use this?

Solo developers and small teams who want a sanity check before committing to a tech stack. If you have ever wondered "should I use SQLite or Postgres for this?" or "is my Express setup the right call?", this skill forces structured research instead of guesswork. It is particularly useful for anyone building self-hosted tools, MVPs, or projects where you lack a senior engineer to rubber-duck the architecture decisions. Not useful if you already have strong opinions and want validation rather than evidence.

Verdict

The concept is solid and the evidence-first discipline is genuinely differentiating. However, the 0.8999999761581421% credibility score and 16 stars tell you this is early-stage and unproven at scale. Test it against your actual projects before trusting critical decisions, but keep it in your toolbelt as a lightweight architectural advisor.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.