Believe Security Documentation
Learn how to analyze security of Solana programs with Believe Security
Introduction
Believe Security is a powerful, GPT-powered analysis tool for Solana programs. It helps developers and security analysts identify potential security vulnerabilities in Solana programs through static analysis and AI-based pattern matching.
Unlike traditional static analyzers, Believe Security leverages advanced language models to understand code semantics and identify complex vulnerability patterns that might be missed by conventional tools.
Most analyses complete in under 60 seconds, providing rapid feedback during development.
Detects common Solana-specific vulnerabilities, logical flaws, and best practice violations.
Receive detailed vulnerability reports with suggestions for remediation.
Quick Start
Getting started with Believe Security is easy. You can analyze your Solana programs in three ways:
Point Believe Security to a public GitHub repository containing your Solana program.
Learn more →Analyze deployed programs by providing a Solana program ID.
Learn more →Upload a ZIP file containing your Solana program source code.
Learn more →Analysis Depth Options
Believe Security offers three levels of analysis depth:
- Quick scan: A rapid analysis that focuses on common vulnerabilities. Perfect for continuous integration workflows or quick development feedback.
- Full: A comprehensive analysis that digs deeper into the codebase and performs more thorough checks. Recommended for pre-deployment review.
- Continuous: Ongoing monitoring and analysis of your program. Ideal for production deployments that you want to keep secure over time.
Vulnerability Categories
Critical Vulnerabilities
Issues that could lead to complete program compromise or fund loss, such as missing signer checks or validation bypasses.
High Severity
Vulnerabilities that may lead to significant security impact or partial fund loss, like integer overflows or unchecked arithmetic.
Medium Severity
Issues that impact security but with limited exploitation potential, such as unchecked owners in CPI calls or account validation issues.
Low Severity
Minor security concerns or best practice violations, like missing rent-exempt checks or improper error handling.
Privacy & Security
Your code is private and secure
Believe Security takes your privacy seriously. All uploaded code and analysis data is automatically deleted after 24 hours. Analysis is performed in isolated, secure environments. We do not store or use your code for training AI models or any other purposes besides the requested analysis.