🔍 Read the full analysis: Learn To Explore Software You Didn’t Write With An AI Agent on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Thorsten Meyer’s October 19 guide describes how the open-source Reverse Engineer Anything (REA) project connects AI coding agents to tools for examining existing software. The article uses a QuickBooks Desktop recreation as a case study, while warning that agent findings need verification and that REA is not a sandbox.
Thorsten Meyer published a guide on October 19 to using the open-source Reverse Engineer Anything (REA) project with AI coding agents to inspect how existing software behaves. The guide presents a QuickBooks Desktop investigation as an example of using REA to turn observed program behavior into a specification, while stressing that agent conclusions must be checked and that the tool is not a security sandbox.
REA combines a local Model Context Protocol (MCP) server, a command-line tool and an agent skill that guides investigations. Meyer says it can connect coding agents including Claude Code, Codex, Cursor and Gemini CLI to analysis tools for software such as JavaScript and Electron apps, .NET and Android programs, firmware, databases and websites observed through a browser. Native program analysis may require tools such as Ghidra, Hopper or IDA.
The guide recommends asking a bounded question about one behavior rather than telling an agent to “figure out” an entire program. A useful investigation, Meyer writes, should identify the decision the developer needs to make and ask for stored data, calculations and conditions that change a result. The agent should distinguish observations from inferences and unknowns, and explain a concrete example with inputs and outputs.
As a case study, Meyer points to Ledger Reborn, a local accounting application developed after its author investigated QuickBooks Desktop for Mac 2024. The source article says its author compared the app’s profit-and-loss statement, balance sheet and trial balance against Intuit’s sample company, with results matching account by account to the cent. Meyer describes this as an illustration of using reverse engineering to document business rules, not as evidence that REA automatically produces a reliable replacement for commercial software.
Infographic 01 · Reverse Engineer Anything
Five places to point REA first, and how to tell when the answer is wrong.
REA gives coding agents such as Claude Code and Codex tools to inspect programs, apps and browser sessions. Each investigation needs a finishable question and a check.
npx rea-agents@latest setupThe five investigations
What to ask, and the check that catches a wrong answer.
- 1
Ask a question you can finish
One behaviour, one decision it informs. Example: how a partial payment changes an invoice balance.
It’s wrong ifThe agent can't walk through one concrete example with real inputs and outputs. - 2
Get REA connected and confirmed
Run the guided setup, then doctor, then call binary_session inside the agent before any real target.
It’s wrong ifAnswers arrive without a single REA tool call: you're reading the model's guesses. - 3
Turn a business rule into a spec
Inputs, stored fields, formula, edge cases, plus test cases checked against the original's own output.
It’s wrong ifYour version and the original disagree on a test case, even by one cent. - 4
See what has to survive a copy
Find what an app writes to the clipboard and what it keeps locally, then design your own copy and paste.
It’s wrong ifThe explanation doesn't predict what lands in a plain text editor, an email and your app. - 5
Your own app, your equipment, an old program
Start with software you own or rely on. Ask what it sends, stores or calculates when you do one thing.
It’s wrong ifIt only works for the example you watched. Test with a new file or a read-only command first.
Sources: REA on GitHub (stars via GitHub API, 19 Oct 2026); REA installation guide; REA security policy.
Powered by Thorsten Meyer AI
From Software Behavior to Specifications
Many software projects depend on behavior that is visible to users but poorly documented: how a payment changes a balance, how an inventory adjustment affects accounts, or which conditions alter a report. REA’s approach, as described by Meyer, offers developers a way to investigate those rules before building an alternative or maintaining a compatible system. That can help replace guesses about what a screen does with a testable account of what the program stores and calculates.
The example also sets a useful standard for AI-assisted software work. A plausible explanation is not enough; the agent should show evidence and walk through a specific case. Meyer’s guide says that if an agent cannot explain a complete example, or answers without calling REA, its output may be a model-generated guess rather than an analysis result. That distinction matters when software behavior affects financial records or other consequential data.
software reverse engineering tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How REA Connects to Coding Agents
The article describes REA as an MIT-licensed open-source project hosted on GitHub. It reports that version 6.7.0 was released October 18, 2026, and that the project had about 325,000 stars and 82,000 forks as of October 19. Those figures are the source article’s snapshot; it does not provide a measurement method or a longer-term comparison.
According to Meyer, setup uses a guided installer and requires Node.js 22.19 or newer. Some forms of native analysis also require a disassembler and its dependencies. The guide recommends checking installation files with REA’s diagnostic command, then confirming inside the agent that its tools actually loaded. Meyer notes that the diagnostic checks files and prerequisites but does not verify a live connection to the agent.
The guide’s accounting example focused on particular questions in the QuickBooks Desktop for Mac 2024 trial, including inventory arithmetic, budget signs by account type, reconciliation balances and cash-basis report paths. Meyer says the investigation used REA 6.4.0 and Ghidra. This is a narrower and more verifiable task than attempting to reconstruct an entire application.
“This is not a sandbox.”
— Reverse Engineer Anything project documentation, as quoted by Meyer
As an affiliate, we earn on qualifying purchases.
Limits of the Reported Results
The article’s QuickBooks comparison is a case study reported by Meyer; the supplied material does not include independent verification, a full test protocol or enough detail to assess every accounting rule tested. Matching three statements for the sample company does not establish that the recreated app matches QuickBooks across all workflows, versions or data.
REA can run analysis locally, but Meyer cautions that findings are passed to the connected agent and may also reach the model provider under that provider’s data policy. The source does not specify the policies of every supported agent or provider. It also does not establish how often agent-generated interpretations are accurate across different software and analysis tools. The project’s maintainer is identified as anonymous in the article.
disassembler and decompiler software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Verify the Agent Before Investigating
Meyer’s immediate recommendation is to confirm that setup succeeded before asking substantive questions: check prerequisites, restart the agent if needed, and verify that it can call an REA tool. He advises beginning with a harmless program the user owns, then choosing one behavior and asking the agent to document a complete example with observed, inferred and unknown points separated.
The source article does not announce a forthcoming release or a formal evaluation of REA’s accuracy. For developers considering the tool, the next step is to review its documentation and safety guidance, check what data the connected model provider receives, and independently test any findings before relying on them in software that handles important records.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Reverse Engineer Anything?
REA is an open-source project that connects coding agents to software-analysis tools through an MCP server, command-line utilities and an agent skill, according to Meyer’s guide.
Does REA make an AI agent’s conclusions reliable?
No accuracy guarantee is established in the article. Meyer recommends requiring evidence and a concrete example, separating observations from inferences, and checking the agent’s findings rather than treating its answer as authoritative.
Does REA run entirely offline?
The analysis tools run locally, Meyer says, but results are sent to the connected agent and may also be sent to its model provider. The applicable data handling depends on that provider’s policy.
Is it safe to analyze an unknown program with REA?
The project warns that REA is not a sandbox. Analysis tools run with the user’s permissions, so the guide advises treating unknown binaries like other untrusted files.
What did the QuickBooks case study show?
Meyer says a developer used REA to investigate selected QuickBooks Desktop for Mac 2024 accounting behaviors and built Ledger Reborn. The source reports that three financial statements matched Intuit’s sample company account by account to the cent, but this is a reported case study, not independent proof of equivalence across all software features.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
