Hakimi Eiqbal (@hakimieiqbal) on X

What it is

A thread by hakimieiqbal on X that maps reverse-engineering tools to target types and centres on REA (github.com/morluto/rea), an open-source layer that connects coding agents such as Claude Code and Codex to RE tools. REA structures investigations around evidence and stated limitations. It covers native binaries, JavaScript/Electron apps, .NET assemblies, websites and runtime behavior. The thread's rule is "evidence > plausible explanations": an agent's convincing answer is not a verified reconstruction.

What to do

REA setup is one command:

npx rea-agents@latest setup

Review the proposed changes, approve them, then restart the agent.

The thread names these MCP bridges for agent-driven analysis:

• IDA Pro MCP (mrexodia)
• GhidrAssistMCP (symgraph)
• GhidraMCP (LaurieWired)
• radare2 MCP (radareorg)

Tools by target, as listed:

• Native: Ghidra, radare2, Rizin, Cutter, plus GDB, LLDB, capa, YARA, Detect It Easy.
• .NET: ILSpy, dnSpyEx, dnlib, de4dot.
• Electron/JS: Electron ASAR, Source Map Explorer, AST Explorer, Babel Parser.
• Firmware: Binwalk, EMBA, FACT Core, FirmAE.
• WebAssembly: WABT, Binaryen, wasm-tools.
• Network: Wireshark, Scapy, Zeek, Boofuzz.
• iOS: Frida, OWASP MASTG, DVIA-v2, libimobiledevice.

The suggested workflow starts with software you own or are authorized to examine. Pick one observable behavior, collect evidence, trace the implementation, write a behavior specification, build an independent reconstruction, compare outputs, and record mismatches and unknowns. The loop is observe, trace, hypothesize, test.

The suggested first prompt is: "Investigate this one feature. Show the relevant files, functions, calls or instructions. Separate verified facts from hypotheses. Predict an edge case, test it where authorized, and report what remains unknown."

The suggested progression is Notes export, Electron clipboard, a practice CTF binary, an APK feature, a native calculation, then firmware or protocols. REA has worked examples for the first two (Notes export, Notion clipboard), plus DX-Ball (native sound calculation) and Aegis (APK TOTP behavior), at rea.tools.

Caveats

• REA is open source, but IDA Pro and Hopper still need their own licenses.
• Analysis runs locally, but tool results are sent to the AI agent, so the model provider's data policy applies.
• The thread is a curated list. It gives no benchmarks or comparisons between tools, and the REA examples are only linked, not described in detail.