Reverse Engineering with AI Agents: Introducing REI Skills and the REA Workbench

Description and use

The article outlines a workflow for using REI Skills and the REA Workbench to perform reproducible reverse engineering of binaries, APKs, firmware, and WebAssembly, emphasizing evidence and verifiability.

Problem

AI agents generate plausible explanations but lack evidence tracking, repeatability, and explicit uncertainty needed for reverse engineering tasks.

Outcome

The workflow enables reproducible, evidence‑linked reverse engineering, exemplified by comparing two WebAssembly binaries and distinguishing observed changes from inferred behavior.

Workflow

• Install REI Skills via npx rei-skills
• Add REA skills with npx rei-skills add rei-rea --<target>
• (Optional) Clone REA workbench repo and install Python dependencies
• Run rei-rea inspect <artifact> to generate evidence
• Run rei-rea diff <artifact1> <artifact2> to compare structures
• Use rei-rea doctor or MCP tools to verify evidence and plan further analysis

Evidence

AI coding agents can generate software surprisingly well. But what happens when the source code is missing?

Source

REI Skills is an open‑source registry and management layer for engineering workflows used by AI coding agents.

Source

REI REA (Reverse Engineering Workbench) integrates 12 specialized agent skills.

Source

The optional Python workbench provides the actual local analysis capabilities.

Source

The repository's recorded demonstration identifies two structural additions: New imported function: env.audit_log_frame and New exported function: validate_packet_header.

Source