Ghidra logo

Ghidra

CommunityPopular
mitsuhiko
ghidra

Reverse engineer binaries using Ghidra's headless analyzer. Decompile executables, extract functions, strings, symbols, and analyze call graphs without GUI.

Overview

Publishermitsuhiko
Repositoryagent-stuff
Skill nameghidra
Stars
3.1K
Forks
224
Bundled files
8
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 8 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by mitsuhiko on GitHub. Read the source before you install it.

Installation

Install the Ghidra AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/mitsuhiko/agent-stuff.git /tmp/agent-stuff
mkdir -p .claude/skills
cp -r /tmp/agent-stuff/skills/ghidra .claude/skills/ghidra
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ghidra in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Ghidra on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Ghidra is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Ghidra Headless Analysis Skill

Perform automated reverse engineering using Ghidra's analyzeHeadless tool. Import binaries, run analysis, decompile to C code, and extract useful information.

Quick Reference

TaskCommand
Full analysis with all exportsghidra-analyze.sh -s ExportAll.java -o ./output binary
Decompile to C codeghidra-analyze.sh -s ExportDecompiled.java -o ./output binary
List functionsghidra-analyze.sh -s ExportFunctions.java -o ./output binary
Extract stringsghidra-analyze.sh -s ExportStrings.java -o ./output binary
Get call graphghidra-analyze.sh -s ExportCalls.java -o ./output binary
Export symbolsghidra-analyze.sh -s ExportSymbols.java -o ./output binary
Find Ghidra pathfind-ghidra.sh

Prerequisites

  • Ghidra must be installed. On macOS: brew install --cask ghidra
  • Java (OpenJDK 17+) must be available

The skill automatically locates Ghidra in common installation paths. Set GHIDRA_HOME environment variable if Ghidra is installed in a non-standard location.


Main Wrapper Script

bash
./scripts/ghidra-analyze.sh [options] <binary>

Wrapper that handles project creation/cleanup and provides a simpler interface to analyzeHeadless.

Options:

  • -o, --output <dir> - Output directory for results (default: current dir)
  • -s, --script <name> - Post-analysis script to run (can be repeated)
  • -a, --script-args <args> - Arguments for the last specified script
  • --script-path <path> - Additional script search path
  • -p, --processor <id> - Processor/architecture (e.g., x86:LE:32:default)
  • -c, --cspec <id> - Compiler spec (e.g., gcc, windows)
  • --no-analysis - Skip auto-analysis (faster, but less info)
  • --timeout <seconds> - Analysis timeout per file
  • --keep-project - Keep the Ghidra project after analysis
  • --project-dir <dir> - Directory for Ghidra project (default: /tmp)
  • --project-name <name> - Project name (default: auto-generated)
  • -v, --verbose - Verbose output

Built-in Export Scripts

ExportAll.java

Comprehensive export - runs all other exports and creates a summary. Best for initial analysis.

Output files:

  • {name}_summary.txt - Overview: architecture, memory sections, function counts
  • {name}_decompiled.c - All functions decompiled to C
  • {name}_functions.json - Function list with signatures and calls
  • {name}_strings.txt - All strings found
  • {name}_interesting.txt - Functions matching security-relevant patterns
bash
./scripts/ghidra-analyze.sh -s ExportAll.java -o ./analysis firmware.bin

ExportDecompiled.java

Decompile all functions to C pseudocode.

Output: {name}_decompiled.c

bash
./scripts/ghidra-analyze.sh -s ExportDecompiled.java -o ./output program.exe

ExportFunctions.java

Export function list as JSON with addresses, signatures, parameters, and call relationships.

Output: {name}_functions.json

json
{
  "program": "example.exe",
  "architecture": "x86",
  "functions": [
    {
      "name": "main",
      "address": "0x00401000",
      "size": 256,
      "signature": "int main(int argc, char **argv)",
      "returnType": "int",
      "callingConvention": "cdecl",
      "isExternal": false,
      "parameters": [{"name": "argc", "type": "int"}, ...],
      "calls": ["printf", "malloc", "process_data"],
      "calledBy": ["_start"]
    }
  ]
}

ExportStrings.java

Extract all strings (ASCII, Unicode) with addresses.

Output: {name}_strings.json

bash
./scripts/ghidra-analyze.sh -s ExportStrings.java -o ./output malware.exe

ExportCalls.java

Export function call graph showing caller/callee relationships.

Output: {name}_calls.json

Includes:

  • Full call graph
  • Potential entry points (functions with no callers)
  • Most frequently called functions

ExportSymbols.java

Export all symbols: imports, exports, and internal symbols.

Output: {name}_symbols.json


Common Workflows

Analyze an Unknown Binary

bash
# Create output directory
mkdir -p ./analysis

# Run comprehensive analysis
./scripts/ghidra-analyze.sh -s ExportAll.java -o ./analysis unknown_binary

# Review the summary first
cat ./analysis/unknown_binary_summary.txt

# Look at interesting patterns (crypto, network, dangerous functions)
cat ./analysis/unknown_binary_interesting.txt

# Check specific decompiled functions
grep -A 50 "encrypt" ./analysis/unknown_binary_decompiled.c

Analyze Firmware

bash
# Specify ARM architecture for firmware
./scripts/ghidra-analyze.sh \
    -p "ARM:LE:32:v7" \
    -s ExportAll.java \
    -o ./firmware_analysis \
    firmware.bin

Quick Function Listing

bash
# Just get function names and addresses (faster)
./scripts/ghidra-analyze.sh --no-analysis -s ExportFunctions.java -o . program

# Parse with jq
cat program_functions.json | jq '.functions[] | "\(.address): \(.name)"'

Find Specific Patterns

bash
# After running ExportDecompiled, search for patterns
grep -n "password\|secret\|key" output_decompiled.c
grep -n "strcpy\|sprintf\|gets" output_decompiled.c

Analyze Multiple Binaries

bash
for bin in ./samples/*; do
    name=$(basename "$bin")
    ./scripts/ghidra-analyze.sh -s ExportAll.java -o "./results/$name" "$bin"
done

Architecture/Processor IDs

Common processor IDs for the -p option:

ArchitectureProcessor ID
x86 32-bitx86:LE:32:default
x86 64-bitx86:LE:64:default
ARM 32-bitARM:LE:32:v7
ARM 64-bitAARCH64:LE:64:v8A
MIPS 32-bitMIPS:BE:32:default or MIPS:LE:32:default
PowerPCPowerPC:BE:32:default

Find all available processors:

bash
ls "$(dirname $(./scripts/find-ghidra.sh))/../Ghidra/Processors/"

Troubleshooting

Ghidra Not Found

bash
# Check if Ghidra is installed
./scripts/find-ghidra.sh

# Set GHIDRA_HOME if in non-standard location
export GHIDRA_HOME=/path/to/ghidra_11.x_PUBLIC
./scripts/ghidra-analyze.sh ...

Analysis Takes Too Long

bash
# Set a timeout (seconds)
./scripts/ghidra-analyze.sh --timeout 300 -s ExportAll.java binary

# Skip analysis for quick export
./scripts/ghidra-analyze.sh --no-analysis -s ExportSymbols.java binary

Out of Memory

Edit the analyzeHeadless script or set:

bash
export MAXMEM=4G

Wrong Architecture Detected

Explicitly specify the processor:

bash
./scripts/ghidra-analyze.sh -p "ARM:LE:32:v7" -s ExportAll.java firmware.bin

Tips

  1. Start with ExportAll.java - It gives you everything and the summary helps orient you
  2. Check the interesting.txt file - It highlights security-relevant functions automatically
  3. Use jq for JSON parsing - The JSON exports are designed to be machine-readable
  4. Decompilation isn't perfect - Use it as a guide, cross-reference with disassembly
  5. Large binaries take time - Use --timeout and consider --no-analysis for quick scans

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Ghidra AI skill do?

Reverse engineer binaries using Ghidra's headless analyzer. Decompile executables, extract functions, strings, symbols, and analyze call graphs without GUI.

Why use Ghidra on TypingMind?

Because you install it once and use it with any model. Ghidra is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Ghidra in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/mitsuhiko/agent-stuff/tree/main/skills/ghidra. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Ghidra?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Ghidra?

As many as you like. As long as a model supports skills, you can use Ghidra with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Ghidra AI skill free?

Yes. It is published on GitHub by mitsuhiko under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇