Security Testing / WebAssembly

Bug and vulnerability discovery in WebAssembly runtimes

A differential testing framework for WebAssembly runtimes that compares behavior across Wasmtime, Wasmer, and V8.

Project objective

The goal was to build an end-to-end differential testing framework that executes the same WebAssembly program across multiple runtimes, captures execution behavior, normalizes runtime-specific outputs, and detects behavioral mismatches.

What the framework captures

• Runtime outputs including stdout, stderr, exit codes, and execution time.
• Normalized outcomes such as ok, trap, validation_error, and runtime_error.
• Structured artifacts including summary.json, mismatches.json, and metrics.md.
• Compiler-generated corpora by compiling C programs into WebAssembly with clang.

Why differential testing works

If multiple runtimes receive the same input but produce different categories of behavior, that mismatch becomes a signal worth investigating. The value is not just running tests; it is creating a pipeline where mismatches can be minimized, reproduced, and analyzed.

This kind of security engineering sits at the boundary of systems, testing, and developer tooling: build a harness, generate realistic inputs, compare behavior, and package results so bugs become actionable.

Deep-dive artifact: view the WebAssembly vulnerability testing PDF.

• WebAssembly
• Security Testing
• Differential Testing
• Wasmtime
• Wasmer
• V8
• Fuzzing
• Runtime Analysis