picofuzz

Picofuzz

A lightweight fuzz-like tool that sends prepared messages using the Fuzz Proto (V1) and collects statistics.

Overview

Picofuzz is designed to test Typeberry nodes by sending pre-prepared binary fuzz messages through socket communication. It establishes a handshake with the target node, processes binary files containing fuzz data, and measures response times while collecting statistics.

Usage

# Run directly with tsx
npx tsx index.ts [options] <directory> <socket>

# Using npm script
npm start [options] <directory> <socket>

# Using Docker
docker build -t picofuzz .
docker run picofuzz [options] <directory> <socket>

Options

• -f, --flavour <spec>: JAM spec: tiny | full (default: tiny)
• -r, --repeat <count>: Number of repetitions (default: 1)
• -s, --stats <file>: Append aggregated stats to a CSV file
• -h, --help: Show help

Parameters

• <directory>: Directory containing .bin files with fuzz data
• <socket>: Socket path or address for communication with the target node

Example

# Run picofuzz on data directory, connect to local socket, repeat 3 times
npm start -r 3 ./fuzz-data /tmp/typeberry.sock

# Run with full JAM spec and save stats to CSV
npm start -f full -s results.csv ./fuzz-data /tmp/typeberry.sock

Testing typeberry with standard w3f-conformance data

# Start typeberry fuzz target first.
npx @typeberry/jam fuzz-target

# Select and execute some tests (data is available via submodules)
npm start -r 3 ../picofuzz-stf-data/picofuzz-data/safrole /tmp/jam_target.sock

How It Works

1. Handshake: Establishes connection with target node using PeerInfo exchange
2. File Processing: Reads all .bin files from the specified directory
3. Message Sending: For each file, decodes the message and sends it to the target
4. Response Handling: Receives and decodes responses from the target node
5. Statistics: Collects timing and success/failure statistics
6. Reporting: Outputs comprehensive statistics at the end of the run

Dependencies

• @typeberry/lib - Core typeberry protocol implementation

Building Docker Image

# Build the Docker image
docker build -t picofuzz .

# Run with mounted data directory
docker run -v /path/to/fuzz-data:/data picofuzz /data /tmp/socket

Statistics

Picofuzz tracks response times for each request

Statistics are printed at the end of each run for performance analysis and debugging.