Mishegos

Tool

First seen 6/9/2026

Last seen 7/1/2026

Evidence 6 chunks

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

RELATIONSHIPS

21 connections

Differential Fuzzing implements → 100% 3e

MISHEGOS implements differential fuzzing to discover decoding discrepancies.

x86_64 instruction decoding evaluates → 100% 3e

MISHEGOS evaluates x86-64 instruction decoders for errors.

Zydis evaluates → 100% 2e

Mishegos evaluates Zydis as one of its fuzzing worker targets.

cohort output uses → 95% 2e

Mishegos organizes decoder outputs into cohorts for analysis.

libbfd evaluates → 100% 2e

Mishegos evaluates libbfd as one of its fuzzing worker targets.

XED evaluates → 100% 2e

Mishegos evaluates XED as one of its fuzzing worker targets.

decoder discrepancy introduces → 95% 2e

Mishegos discovers and reports decoder discrepancies between different x86_64 decoder implementations.

mishegos worker ← part of 95% 2e

Mishegos is composed of worker processes, each wrapping a decoder target with a defined ABI.

Difference Oracle uses → 1e

MISHEGOS uses a difference oracle to detect erroneous behavior between decoders.

sliding mutation engine implements → 100% 1e

Mishegos uses the sliding mutation engine to generate x86_64 instruction candidates by extracting windows of up to 15 bytes.

shared memory slot uses → 95% 1e

Mishegos uses shared memory slots to pass inputs to workers and collect outputs.

Sliding Window Fuzzing Strategy uses → 1e

MISHEGOS uses the sliding window fuzzing strategy for variable-length instruction architectures.

mishmat uses → 95% 1e

Mishegos uses mishmat to visualize analysis output as HTML tables.

Decoding Discrepancy introduces → 1e

MISHEGOS surfaces decoding discrepancies between instruction decoders.

passes.yml ← part of 90% 1e

passes.yml is the configuration file for analysis passes in Mishegos.

Trail of Bits authored by → 95% 1e

Mishegos is open-sourced by Trail of Bits.

Differential Analysis of x86-64 Instruction Decoders ← introduces 1e

The paper introduces MISHEGOS as a novel differential fuzzer.

worker process ABI uses → 95% 1e

Each fuzzing target in Mishegos is wrapped into a worker process with a well-defined ABI.

analysis pass uses → 90% 1e

Mishegos's analysis tool organizes filters into passes to process output cohorts.

analysis filter uses → 90% 1e

Mishegos uses analysis filters as small scripts to process and filter cohort output.

input minimization uses → 90% 1e

Mishegos includes a minimize-input filter that reduces instruction candidates to the minimal relevant length.