Welcome to EMYAC - the Extensive Malware YARA Accuracy Competition. Here, you'll learn how to write effective YARA rules to detect malware families across real sample sets. Follow this guide to get started.
YARA is a tool used to identify and classify malware based on textual or binary patterns. It helps reverse engineers and threat analysts detect malware by creating custom rules based on known indicators.
YARA-X is the modern, performance-focused evolution of YARA, with improved capabilities and architecture.
YARA Documentation | YARA-X GitHub
Select a malware family you're interested in learning about. You’ll be given access to a subset of sample files associated with that malware.
AgentTesla
or Formbook
before tackling more advanced ones like CobaltStrike
.After selecting a family, download the samples. Each challenge will provide a curated collection of binaries for analysis.
Important: These are real malware samples. Read the next step carefully.
Always work in a secure environment. Use virtual machines or sandboxes such as:
Now it's time to craft your YARA rule. Focus on creating strong, relevant conditions based on strings, hex patterns, file sizes, or imported functions.
meta: author = "yourname"
to be credited.Begin by testing your rule on your local system to ensure your rule matches all samples.
yara your_rule.yar /path/to/samples
Check which files were matched and adjust your rule to improve it.
yara your_rule.yar /path/to/samples -s
Once you’re happy with your results, submit your rule for scoring. You'll get feedback on detection accuracy, precision, recall, rule performance and the exact number of false positives/false negatives.
Use the feedback to refine your rule:
Learn more about precision and recall
Dive in, learn through doing, and sharpen your skills in writing YARA rules. Your progress and results will be reflected in the leaderboard and scoring system.
This competition would not be possible without Lena, creator of Malmons, and the whole Malware Village team. Infrastructure was kindly provided by GITC and samples are courtesy of vxunderground and VirusTotal.