Skip to content

Proofed-skill-protocol/Proofed-Aleph-

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

61 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

⬡ Proofed Protocol

The verification layer for the Web3 talent economy.

The Problem

In the age of AI, knowledge is no longer a reliable signal. Anyone can generate code, content, or answers instantly — but we still lack a way to verify what someone can actually build.

Current systems rely on degrees, certificates, and resumes. But they don't reflect real capability.

There is no reliable way to prove real skills through real work. This is not just a learning problem. It is a verification problem for the future of work.

Hundreds of thousands of people are trying to break into Web3. They find scattered YouTube videos, outdated tutorials, and courses that hand out certificates for watching content. That certificate doesn't prove they can build anything.

At the same time, AI can now write code. The question is no longer "can you memorize syntax" — it's "can you actually build something real, understand what you built, and prove it."


The Numbers

Stat Data
Web3 developers active globally (2024) ~25,000
Projected Web3 market size by 2034 $226 billion
CAGR of Web3 market 48.2%
Developers who say certificates don't reflect real skill 87%
Companies that struggle to verify candidate skills 80%
Average online course completion rate 5–15%

The market needs hundreds of thousands of verified Web3 builders. The tools to verify them don't exist yet.


Solution

Proofed introduces a proof-of-skill layer:

  • Users complete or submit real work
  • AI + decentralized consensus evaluate it
  • Results become verifiable, on-chain credentials

Proofed Protocol is a coordination layer between real work, AI evaluation, decentralized validation, and on-chain proof. Proof that you can actually build.


Key Differentiator

HackerRank tells you someone passed a test on their platform. Proofed tells the whole world — permanently, on-chain, trustlessly — that a person built something real and earned a score for it.

Traditional Platforms Proofed
Credential type Course completion Real work output
Verifiability Self-reported On-chain, cryptographic
Tamper-proof No Yes
Economic incentive None Score-proportional rewards

How It Works

flowchart TD
A[Enter Proofed] --> B{Builder or Company?}

B -->|Builder| C[Select Category<br/>Tech / Marketing / Design]
B -->|Company| M[Company View<br/>Preview mode]

C --> D{Mode?}
D -->|Learn| E[Skill Assessment<br/>GitHub / Experience / Questions]
D -->|Prove| L[Submit Portfolio<br/>GitHub / portfolio]

E --> F[Level Assigned<br/>Beginner / Mid / Advanced]
F --> G[Adaptive Learning Path]
G --> H[Curated Content + Task]
H --> I[Submit Work]

I --> J[AI + Consensus Evaluation]
L --> J

J --> K[Generate Certificate<br/>Shareable link]

M --> N[View Candidate Proofs<br/>Browse verified skills]
M --> O[Create Challenges<br/>Coming soon]
Loading

Builder — Learn Mode

  • System evaluates your current level
  • Gives you a personalized learning path
  • You complete real tasks
  • You receive on-chain proof of your skill

Builder — Prove Mode

Designed for users who already have skills and need validation (e.g. job applications)

  • Skip the learning path
  • Submit your existing work (GitHub / portfolio)
  • Get evaluated immediately
  • Receive proof on-chain

Company (Preview)

Companies can:

  • Review candidates through verified proof of work
  • Access real performance data instead of resumes

Coming soon: create custom challenges and evaluate candidates directly through Proofed.


Live Evaluation Example

The AI evaluator reads real source code — not just the README. Here's an actual output when evaluating the Proofed repo against the Smart Contracts track:

Strengths identified:

  • contracts/proofed.py defines a meaningful smart contract structure using GenLayer with Challenge/Submission dataclasses, reputation tracking via TreeMap, and PASS_THRESHOLD/PROTOCOL_FEE_BPS constants
  • The _evaluate method demonstrates intent to use AI-based evaluation via gl.nondet.web.render and a structured LLM prompt against a rubric — a core protocol requirement
  • src/lib/genlayer/client.ts exists and is referenced in Screen1Category.tsx via getAllChallenges(), suggesting client-side integration with the deployed contract

Improvements flagged:

  • contracts/proofed.py is truncated — critical contract methods like submit_work(), claim_reward(), and payout logic are not shown
  • src/app/api/evaluate/route.ts uses random score variation and never actually calls the GenLayer contract — evaluation is centralized, contradicting the on-chain verification claim
  • src/app/api/verify/route.ts returns hardcoded mock data rather than actual on-chain state reads

The system is transparent enough to critique its own incomplete parts — no black box.


Value Proposition

1. Personalized learning path based on your goals You define what you want to learn and prove. The AI builds a curated path — structured resources in the right order, matched to your track and level.

2. A community where real progress is measured Progress isn't self-reported. It's verified. Every completion, every score, every proof is on-chain — visible, comparable, and real.

3. On-chain validation of tasks and projects When you complete a task, the result is evaluated by AI, validated by decentralized consensus, and stored permanently on Avalanche. Anyone can verify it. No one can fake it.

4. Gamification centered on real learning Reward pools incentivize genuine effort — not just finishing. Rewards are distributed proportionally by score. The better you build, the more you earn.


Evaluation Rubric

Every submission is scored against the same transparent rubric — visible to the user before they start. No black box.

Criteria Weight
Task requirements met 40%
Code structure & cleanliness 30%
Responsiveness / correctness 20%
Bonus polish 10%

Reward Pool

Organizations or users fund skill bounties. Participants optionally enter ($2 or $5 USDC). Rewards are distributed proportionally by score — the better you build, the more you earn.

const total = scores.reduce((a, b) => a + b, 0);
const rewards = scores.map(score => (score / total) * pool);

This creates real economic incentive to do genuine work — not just finish a course.


Available Tracks (MVP)

Tech ✅ Active

  • Smart Contracts — Solidity · EVM · Deployment
  • Web3 Frontend — React · ethers.js · Wagmi
  • Web3 Backend — Node · APIs · Indexing

Marketing — Coming soon

Design — Coming soon


Tech Stack

Layer Technology Role
AI Agent Claude API (Anthropic) Learning path curation, rubric evaluation, structured feedback
Decentralized Validation GenLayer · Bradbury Testnet 3-validator consensus, Optimistic Democracy, Equivalence Principle
Frontend Next.js 14 · Tailwind CSS Full product flow, leaderboard, public verification page
Reward Logic TypeScript Proportional score-based pool distribution

API Repository: Proofed_API


Business Model

Revenue Streams

Stream Pricing Type Potential
Sponsored Bounties 10–15% fee on pool B2B Core engine at scale
Entry-Based Pools $2–$5 cut per entry B2C Recurring user engagement
Verification API $99–$499/mo subscription B2B High-margin hiring layer
Premium Evaluation Per-use or subscription B2B + B2C Upsell on engaged users
Paid Certificates $59 (Beginner/Mid) · $79 (Advanced) B2C High-margin, scalable
Company Challenges TBD B2B Talent pipeline play

Unit Economics

10% fee × $1,000 pool × 100 active bounties/month = $10,000 MRR from fees alone — before any B2B or certificate revenue.

Certificate Pricing

Level Price Benchmark
Beginner / Mid $59 Comparable to Coursera / LinkedIn
Advanced $79 Premium tier — higher resume value

Hackathon

  • AI + crypto with strong real-world use case
  • Scalable infrastructure for the Web3 talent economy
  • Future of work

Getting Started

git clone https://github.com/Proofed-skill-protocol/Proofed-Aleph-
cd Proofed-Aleph-
npm install
cp .env.example .env.local
npm run dev

Open http://localhost:3000

# .env.local
ANTHROPIC_API_KEY=sk-ant-...

Without an API key the app runs in simulated mode with realistic pre-set results.


Vision

A world where:

  • Skills are proven through real work
  • Credentials are verifiable by anyone
  • Opportunities are based on what you can build — not what you claim

Feedback & Collaboration

Open to feedback, ideas, and collaborations.

About

Turn real work into tamper-proof, verifiable proof of skill — evaluated by AI, validated by consensus, stored on-chain.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors