<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk" -->

---
title: LLM assisted reverse engineering | daily.dev
description: A longtime reverse engineer describes using LLM agents to speed up hacking hardware and software. The workflow pairs DeepSeek v4.1 Flash with Ghidra via an MCP...
canonical: https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: LLM assisted reverse engineering | daily.dev
og:description: A longtime reverse engineer describes using LLM agents to speed up hacking hardware and software. The workflow pairs DeepSeek v4.1 Flash with Ghidra via an MCP...
og:url: https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk
og:image: https://api.daily.dev/og/posts/Q75olVwVK.png
og:image:alt: LLM assisted reverse engineering
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM assisted reverse engineering

**[Hacker News](https://daily.dev/sources/hn)** · 7 min read · 0 upvotes · 0 comments

## Summary

A longtime reverse engineer describes using LLM agents to speed up hacking hardware and software. The workflow pairs DeepSeek v4.1 Flash with Ghidra via an MCP bridge, plus radare2 and kuna for faster CLI disassembly. Agents are told to log findings to MD files and follow a 'hypothesis debugging' loop: hypothesize, design an experiment, record results, repeat. Rust is used for final implementations because its compiler gives strict feedback to agents, while Python with uv handles quick prototyping. Projects completed in six months include reverse-engineered firmware for the Teenage Engineering OP-Z, Akai MPC, Arturia MiniFreak, and a hacked Alphasmart NEO device.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.slowchronicle.com/posts/llm-assisted-reverse-engineering>

## Questions this post answers

### How can I use an MCP bridge with Ghidra for AI-assisted reverse engineering?

An MCP bridge connects Ghidra's disassembler and decompiler to an LLM agent so the agent can query and manipulate binary analysis directly. Combined with a dedicated tooling skill describing how to use the bridge, agents handle disassembly and decompilation far more effectively instead of giving up when they get lost.

_Reverse engineers wiring up agent tooling can track workflow write-ups like this on daily.dev._

### What is the hypothesis debugging technique for unsticking an AI coding agent?

It's a scientific-method loop for agents stuck on a hard problem: write a hypothesis about why something works into a saved CSV table, design an experiment (a probe, logging, breakpoint) that definitively proves or disproves it, record the result, and repeat. This stops the agent from endlessly guessing or looping on random attempts.

_Developers tuning agent workflows for stubborn debugging sessions follow techniques like this on daily.dev._

### Why use Rust instead of Python when implementing a reverse-engineered hack or firmware mod?

Rust is preferred for final implementation because its strict compiler gives agents immediate, precise feedback on memory and syntax issues during iteration, which tightens the agent's feedback loop. Python with uv is still used for quick prototyping, probes, and single-file inline-dependency scripts, despite Rust's slower compile times.

_Developers weighing Rust against Python for agent-driven implementation work can follow comparisons like this on daily.dev._

## Community take

How the wider developer community reacted, aggregated from 1 discussion and 41 comments across hackernews (as of 2026-11-19).

**TL;DR:** Developers share enthusiasm for using LLMs to speed up reverse engineering of firmware, apps, and protocols, while a large side debate focuses on legal/ethical boundaries, model guardrails, and whether offloading the work undermines the skill-building value of RE.

**Sentiment:** 40% positive · 25% mixed · 35% skeptical

**The case for**

- Several people report successfully using LLMs to reverse engineer plugins, firmware, cloud/IoT protocols, and APKs in hours instead of much longer manual efforts.
- Hypothesis-driven, experiment-logging workflows are seen as effectively curbing LLMs' tendency to guess randomly or loop.
- Tasking LLMs with writing emulation harnesses (e.g. Unicorn) for bootroms/firmware is called out as a particularly strong use case.
- Some find that reviewing the agent's logged findings afterward still lets them learn from the process.

**The pushback**

- Some models (especially certain Claude/Anthropic versions) frequently refuse or trigger guardrails mid-task, forcing people to switch to less restricted models like DeepSeek.
- There's concern that using AI-generated code for RE/patents could introduce legal liability or shipped infringing code that nobody actually reviews.
- Several argue that offloading the RE process to an agent removes the skill-building and satisfaction that come from doing the work manually.
- Debate over the ethics of using LLMs to crack paid apps/DRM, with some calling it straightforward piracy dressed up in justification.

**By community**

- hackernews (heated): Discussion splits between practical enthusiasm for LLM-assisted RE workflows and a sharp ethical/legal argument over piracy, guardrail-bypassing, and code-infringement liability.

**Hottest debate:** Whether using LLMs to reverse engineer and unlock paid apps constitutes legitimate reuse/privacy improvement or is simply theft dressed up as principle.

**Open questions**

- How will closed model providers' guardrails against reverse engineering evolve, and will they become impossible to work around?
- What happens legally when AI-generated code used in RE or reimplementation infringes on existing patents or copyrighted implementations?
- Does relying on agents for RE undermine the learning and satisfaction that come from doing it manually?

**Highlights**

> > I guess Android apps are just easy to reverse engineer or something. Android apps are indeed easier than average to reverse engineer as they are effectively Dalvik (JVM-like) bytecode wrapped in a zip file format, and Dalvik/JVM style bytecode is among the easiest compiled/shipped executable formats to decompile to a usable source code representation. More often than not the only protection they have is symbol/function name obfuscation, which makes things a bit tedious for human reverse engineering but which doesn't slow LLMs down much.
> — [georgemcbay on hackernews](https://news.ycombinator.com/item?id=50556537)

> > This is variation of scientific method and it significantly helps agents when they are getting stuck on some problem. Stops it going in circles. By forcing it to do definitive experiments. No endless guessing and trying random things. These "guessing loops" are my main annoyance when REing with Claude (mostly Opus 4.8, due to the restrictions on newer models). On harder/larger targets, I have to constantly babysit it and tell it to "stop guessing". For truly black-box reverse engineering, yes, hypothesis-testing is the way to go (it's all you can do). But in most cases all the information is available to arrive at an answer deterministically, it just needs to be worked through meticulously. As an experienced human I have strong intuitions and I can work much faster under a "guess and check" model, but I prefer to use Claude as an "unstoppable RE juggernaut" with superhuman patience and determination, tracing method invocations all the way from A to distant subsystem B (for example). Another big win for hardware RE is from tasking the LLM with writing unicorn emulation harnesses, e.g. for bootroms and firmwares. It's pretty good at the classic "figure out which MMIO needs to be stubbed/simulated next to progress the boot" loop, although sometimes it tries to take egregious purpose-defeating shortcuts.
> — [Retr0id on hackernews · 1 comments](https://news.ycombinator.com/item?id=50554878)

> This is usually among the list of reasons FOSS projects / people reject the use of LLM generated code entirely, see for example Codeberg. The answer is, it's almost certainly shipping infringing code, since there's pretty damning examples of this happening. [0] Whether this matters is to be decided. Given that AI technologies are the backbone of the American economy right now, and given that the people running these companies are all very rich and have lots of rich friends, as well as friends in the government, and given that the world of FOSS is generally not full of very rich or very well connected people, I suspect it won't matter in the end. Some kind of carve-out will be created for LLM generated code, if necessary. Or they can just depend on the fact that nobody can afford to sue over it. [0] https://pmc.ncbi.nlm.nih.gov/articles/PMC12377501
> — [komali2 on hackernews · 1 comments](https://news.ycombinator.com/item?id=50555663)

> Yesterday night, i reverse engineered the cloud connection plugin to my cctv camera in about 2 hours with Gemini. Started off with me searching for what Chinese OEM made the product. Then tried out existing GitHub projects to see if it works but they were failing. Spent 15 minutes to write a script that started their desktop software on a different network namespace and ran tcpdump in it. Captured exactly only the things i needed without any noise. Turned out it was using a different server url and credentials than the GitHub projects. Opened Ghidra to inspect the dlls that matched the strings i was looking for. After that, I got lazy and handed over the task to Gemini to find the username and key they use to authenticate with the servers. It took about 30 minutes or an hour. A lot of it was me changing models because the llm servers were busy. It simply wrote a little script that ran through various strings it fetched from the binary to see if it can generate the authentication credentials. Now my Home Assistant will be detecting if my car is parked in the parking lot or not. Life is good.
> — [saidinesh5 on hackernews](https://news.ycombinator.com/item?id=50556237)

> For me, the process is what matters, because it’s through the process that you gain skills, and achieving the final result brings a great deal of satisfaction. When I use AI for reverse engineering, I get neither the former nor the latter.
> — [therepanic on hackernews · 2 comments](https://news.ycombinator.com/item?id=50555496)

**Source threads**

- [hackernews](https://news.ycombinator.com/item?id=50553225) · 128 points · 41 comments

## Similar posts on daily.dev

- [On the Coming Industrialisation of Exploit Generation with LLMs](https://daily.dev/posts/on-the-coming-industrialisation-of-exploit-generation-with-llms-wkt9zuzy1) · Hacker News · 2 upvotes · 0 comments
- [AI Makes Reverse Engineering Mainstream Software & Malware Easy](https://daily.dev/posts/ai-makes-reverse-engineering-mainstream-software-malware-easy-imzni0rfo) · Dark Reading · 0 upvotes · 0 comments
- [LLM-Generated Mythic Agents Enable Disposable Red-Team Tooling](https://daily.dev/posts/llm-generated-mythic-agents-enable-disposable-red-team-tooling-bx5ghnvyo) · Security Boulevard · 0 upvotes · 0 comments

---

Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#rust](https://daily.dev/tags/rust), [#mcp](https://daily.dev/tags/mcp), [#reverse-engineering](https://daily.dev/tags/reverse-engineering)

[View this post on daily.dev](https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"LLM assisted reverse engineering","url":"https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk"},"datePublished":"2026-11-19T01:20:08.037Z","dateModified":"2026-11-19T06:06:19.161Z","description":"A longtime reverse engineer describes using LLM agents to speed up hacking hardware and software. The workflow pairs DeepSeek v4.1 Flash with Ghidra via an MCP...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/66f205331cd961f31652e8e3d9aa068e?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/66f205331cd961f31652e8e3d9aa068e?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,rust,mcp,reverse-engineering","timeRequired":"PT7M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"LLM assisted reverse engineering"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/llm-assisted-reverse-engineering-q75olvwvk#faq","mainEntity":[{"@type":"Question","name":"How can I use an MCP bridge with Ghidra for AI-assisted reverse engineering?","acceptedAnswer":{"@type":"Answer","text":"An MCP bridge connects Ghidra's disassembler and decompiler to an LLM agent so the agent can query and manipulate binary analysis directly. Combined with a dedicated tooling skill describing how to use the bridge, agents handle disassembly and decompilation far more effectively instead of giving up when they get lost. Reverse engineers wiring up agent tooling can track workflow write-ups like this on daily.dev."}},{"@type":"Question","name":"What is the hypothesis debugging technique for unsticking an AI coding agent?","acceptedAnswer":{"@type":"Answer","text":"It's a scientific-method loop for agents stuck on a hard problem: write a hypothesis about why something works into a saved CSV table, design an experiment (a probe, logging, breakpoint) that definitively proves or disproves it, record the result, and repeat. This stops the agent from endlessly guessing or looping on random attempts. Developers tuning agent workflows for stubborn debugging sessions follow techniques like this on daily.dev."}},{"@type":"Question","name":"Why use Rust instead of Python when implementing a reverse-engineered hack or firmware mod?","acceptedAnswer":{"@type":"Answer","text":"Rust is preferred for final implementation because its strict compiler gives agents immediate, precise feedback on memory and syntax issues during iteration, which tightens the agent's feedback loop. Python with uv is still used for quick prototyping, probes, and single-file inline-dependency scripts, despite Rust's slower compile times. Developers weighing Rust against Python for agent-driven implementation work can follow comparisons like this on daily.dev."}}]}
```

