Back to portfolio
ProductIn production2025–Present

CryptoSecure

AI-powered security scanner for TON blockchain smart contracts. Paste a FunC or Tact contract and CryptoSecure (branded TON Guardian) finds the vulnerabilities, explains each one in plain English, and hands back line-by-line fixes — the kind of pre-deployment audit that normally costs $10K–$50K and takes weeks, delivered in under 30 seconds. Built end-to-end as a Next.js dashboard + AI analysis engine + custom vulnerability database.

Gallery

Instant scan

Paste a contract or load a sample — findings return in under 30 seconds with a 0–100 security score.

Plain-English findings

Every vulnerability is explained in readable language with its real-world impact, not just a rule ID.

Actionable fixes

Line-by-line remediation and secure code examples so developers can ship the fix, not just spot the bug.

Product

CryptoSecure is a security tool built for TON developers who need an audit-grade check before deployment, without the audit-grade price tag or timeline. It combines a static analysis engine with an LLM layer so findings stay grounded in the actual contract code and are explained in a way any developer can act on.

Problem

Recent research found 14,995 vulnerabilities across 1,640 TON smart contracts — over 9 bugs per contract on average. Professional audits run $10K–$50K and take 2–4 weeks, putting security out of reach for most independent builders and small teams.

Solution

A free, instant scanner that detects 8+ vulnerability classes, scores each contract 0–100, and translates every finding into plain English with concrete fixes. Core scanning requires no API key; an optional “Hacker Mode” adds deeper LLM analysis with your own OpenAI credentials.

Highlights

  • AI + static analysis engine detecting 8+ vulnerability classes (reentrancy, access control, integer overflow, unchecked return values, and more)
  • 0–100 security score with severity breakdown and per-issue impact
  • Plain-English explanations paired with line-by-line remediation and secure code examples
  • Sub-30-second scans on FunC and Tact contracts
  • Pre-loaded sample contracts for zero-setup evaluation
  • Free core scanning with no API key; optional “Hacker Mode” for deeper LLM analysis
  • Security-score dashboard built with Recharts for at-a-glance results

Specs

Type
AI product / security tool
Role
Full-stack engineer
Company
Independent
Focus
Smart contract vulnerability detection
Period
2025–Present
Chain
TON (FunC / Tact)
Deployment
Vercel
Repository
github.com/the-sniper/crypto_secure

How it works

  1. 1Submit a contract — Paste FunC/Tact code or load a pre-built sample as the input.
  2. 2Analyze — A static analysis engine plus GPT-4o inspect the code against a custom vulnerability database.
  3. 3Score & explain — Findings are graded by severity and rolled into a 0–100 security score with plain-English write-ups.
  4. 4Fix — Each issue comes with line-by-line remediation and secure code examples the developer can apply directly.

Achievements

  • Shipped a full product end-to-end: Next.js dashboard + AI analysis engine + custom vulnerability database
  • Built a detection layer covering 8+ vulnerability classes across FunC and Tact contracts
  • Turned raw findings into a 0–100 security-score dashboard with severity breakdowns and impact explanations
  • Made security accessible — sub-30-second, free scans versus $10K–$50K, multi-week manual audits
  • Designed a dual-mode access model: free static scanning plus optional key-based “Hacker Mode” for deeper LLM analysis
  • Deployed to production on Vercel with a TypeScript codebase (98%+)

Stack

TypeScriptNext.jsReactTailwind CSSRechartsLucide ReactOpenAI GPT-4o
Core StackFrontendData & Machine LearningDesign & DX

Links

Repository