MCP/mcghidra
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: Simplify Docker setup and modernize project structure
Some checks are pending
Build Ghidra Plugin / build (push) Waiting to run
Details

Browse Source 
- Switch from Java to Python scripts (avoids OSGi bundle issues)
- Update pyproject.toml with proper src layout and ruff config
- Add binaries/ and ghidra-src/ to gitignore
- Clean up Module.manifest
This commit is contained in:
Ryan Malloy 2026-01-26 13:50:36 -07:00
parent 88e1fe6ca8
commit a3ad70d302
5 changed files with 254 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
.gitignore vendored
View File

@ -50,3 +50,10 @@ hs_err_pid*
replay_pid*
/__pycache__
# Docker volumes and binaries
/binaries/*
!/binaries/.gitkeep
# Ghidra source (fetched separately)
/ghidra-src/

22
docker/Dockerfile
View File

@ -110,20 +110,16 @@ RUN mkdir -p /opt/ghidra/Ghidra/Extensions \
RUN mkdir -p /projects /binaries /home/ghidra/.ghidra \
&& chown -R ghidra:ghidra /projects /binaries /home/ghidra
# Copy GhydraMCP scripts to the BSim module's scripts directory
# BSim is a working feature module with proper OSGi bundle configuration for scripts
COPY docker/GhydraMCPServer.java /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/
COPY docker/ImportRawARM.java /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/
COPY docker/TestScript.java /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/
# Copy GhydraMCP Python scripts to user scripts directory
# Python/Jython scripts don't require OSGi bundle registration - they work without issue
RUN mkdir -p /home/ghidra/ghidra_scripts
COPY docker/GhydraMCPServer.py /home/ghidra/ghidra_scripts/
COPY docker/ImportRawARM.java /home/ghidra/ghidra_scripts/
# Set proper ownership, permissions, and timestamp to match Ghidra installation
# Ghidra appears to validate scripts by timestamp - newer files may be rejected
RUN chown ghidra:ghidra /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/GhydraMCPServer.java \
/opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/ImportRawARM.java \
/opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/TestScript.java \
&& touch -t 202508261420 /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/GhydraMCPServer.java \
&& touch -t 202508261420 /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/ImportRawARM.java \
&& touch -t 202508261420 /opt/ghidra/Ghidra/Features/BSim/ghidra_scripts/TestScript.java
# Set proper ownership and permissions
RUN chown -R ghidra:ghidra /home/ghidra/ghidra_scripts \
&& chmod 755 /home/ghidra/ghidra_scripts/*.py 2>/dev/null || true \
&& chmod 755 /home/ghidra/ghidra_scripts/*.java 2>/dev/null || true
# Copy entrypoint script (755 so ghidra user can read and execute)
COPY docker/entrypoint.sh /entrypoint.sh

204
docker/ImportRawARM.java Normal file
View File

@ -0,0 +1,204 @@
// Import and analyze raw ARM firmware binary
// This script imports a raw binary file with specified ARM processor and load address
// @author GhydraMCP
// @category Binary.Import
// @keybinding
// @menupath
// @toolbar
import ghidra.app.script.GhidraScript;
import ghidra.app.util.bin.ByteArrayProvider;
import ghidra.app.util.importer.MessageLog;
import ghidra.app.util.opinion.BinaryLoader;
import ghidra.app.util.opinion.LoadSpec;
import ghidra.framework.model.DomainFile;
import ghidra.framework.model.DomainFolder;
import ghidra.program.model.address.Address;
import ghidra.program.model.lang.LanguageCompilerSpecPair;
import ghidra.program.model.listing.Program;
import ghidra.util.task.TaskMonitor;
import java.io.File;
import java.io.IOException;
import java.nio.file.Files;
import java.util.List;
public class ImportRawARM extends GhidraScript {
@Override
public void run() throws Exception {
String[] args = getScriptArgs();
if (args.length < 2) {
println("========================================");
println(" Import Raw ARM Firmware Binary");
println("========================================");
println("");
println("Usage: -postScript ImportRawARM.java <binary_path> <load_address> [<language_id>]");
println("");
println("Arguments:");
println(" binary_path - Path to raw binary file");
println(" load_address - Base address in hex (e.g., 0x00000000)");
println(" language_id - ARM language ID (default: ARM:LE:32:v5t)");
println("");
println("Common ARM Language IDs:");
println(" ARM:LE:32:v4 - ARMv4 little-endian");
println(" ARM:LE:32:v5 - ARMv5 little-endian");
println(" ARM:LE:32:v5t - ARMv5T little-endian (Thumb)");
println(" ARM:LE:32:v6 - ARMv6 little-endian");
println(" ARM:LE:32:v7 - ARMv7 little-endian");
println(" ARM:LE:32:Cortex - ARM Cortex");
println(" ARM:BE:32:v5t - ARMv5T big-endian");
println("");
println("Example:");
println(" analyzeHeadless /projects MyProject \\");
println(" -postScript ImportRawARM.java /binaries/firmware.bin 0x00000000 ARM:LE:32:v5t");
println("");
return;
}
String binaryPath = args[0];
String loadAddressStr = args[1];
String languageIDStr = args.length > 2 ? args[2] : "ARM:LE:32:v5t";
// Parse load address
long loadAddress;
if (loadAddressStr.toLowerCase().startsWith("0x")) {
loadAddress = Long.parseLong(loadAddressStr.substring(2), 16);
} else {
loadAddress = Long.parseLong(loadAddressStr, 16);
}
File binaryFile = new File(binaryPath);
if (!binaryFile.exists()) {
printerr("ERROR: Binary file not found: " + binaryPath);
return;
}
long fileSize = binaryFile.length();
println("========================================");
println(" Importing Raw ARM Firmware");
println("========================================");
println(" Binary: " + binaryFile.getName());
println(" Size: " + fileSize + " bytes (0x" + Long.toHexString(fileSize) + ")");
println(" Load Address: 0x" + String.format("%08X", loadAddress));
println(" Language: " + languageIDStr);
println("========================================");
println("");
try {
// Read the binary file
byte[] bytes = Files.readAllBytes(binaryFile.toPath());
ByteArrayProvider provider = new ByteArrayProvider(bytes);
// Parse language string (format: "ARM:LE:32:v5t")
String[] parts = languageIDStr.split(":");
if (parts.length < 4) {
printerr("ERROR: Invalid language ID format. Expected format: ARCH:ENDIAN:SIZE:VARIANT");
printerr(" Example: ARM:LE:32:v5t");
return;
}
// Create language compiler spec pair
LanguageCompilerSpecPair lcsPair = new LanguageCompilerSpecPair(languageIDStr, "default");
// Use BinaryLoader to import
BinaryLoader loader = new BinaryLoader();
// Create load spec with our language
LoadSpec loadSpec = new LoadSpec(loader, loadAddress, lcsPair, false);
// Get project folder
DomainFolder rootFolder = state.getProject().getProjectData().getRootFolder();
String programName = binaryFile.getName();
println("Creating program: " + programName);
println("Using language: " + languageIDStr);
println("");
// Import using BinaryLoader
MessageLog log = new MessageLog();
List<Program> programs = loader.load(
provider,
programName,
rootFolder,
loadSpec,
List.of(), // options
log,
this, // consumer
monitor
);
if (programs == null || programs.isEmpty()) {
printerr("ERROR: Failed to load binary");
println("");
println("Loader messages:");
println(log.toString());
return;
}
Program program = programs.get(0);
println("Import successful!");
println(" Program: " + program.getName());
println(" Base Address: " + program.getImageBase());
println(" Memory blocks: " + program.getMemory().getBlocks().length);
println("");
// Set entry point at load address
Address entryAddr = program.getAddressFactory().getDefaultAddressSpace().getAddress(loadAddress);
int txId = program.startTransaction("Set Entry Point");
try {
program.getSymbolTable().addExternalEntryPoint(entryAddr);
program.getSymbolTable().createLabel(entryAddr, "entry",
ghidra.program.model.symbol.SourceType.ANALYSIS);
program.endTransaction(txId, true);
println("Entry point set at: " + entryAddr);
println("");
} catch (Exception e) {
program.endTransaction(txId, false);
printerr("Warning: Could not set entry point: " + e.getMessage());
}
// Save the program
try {
program.save("ARM firmware import", monitor);
println("Program saved successfully");
println("");
} catch (Exception e) {
printerr("Warning: Could not save program: " + e.getMessage());
}
// Run analysis
println("Starting auto-analysis...");
println("(This may take a while for large binaries)");
println("");
analyzeAll(program);
println("");
println("========================================");
println(" Import Complete!");
println("========================================");
println(" Program: " + program.getName());
println(" Functions found: " + program.getFunctionManager().getFunctionCount());
println(" Defined data: " + program.getListing().getNumDefinedData());
println("========================================");
println("");
// Set as current program for subsequent scripts
state.setCurrentProgram(program);
} catch (IOException e) {
printerr("ERROR: Failed to read binary file: " + e.getMessage());
e.printStackTrace();
} catch (Exception e) {
printerr("ERROR: Import failed: " + e.getMessage());
e.printStackTrace();
}
}
}

29
pyproject.toml
View File

@ -1,21 +1,42 @@
[project]
name = "ghydramcp"
version = "2025.12.1"
version = "2025.12.3"
description = "AI-assisted reverse engineering bridge: a multi-instance Ghidra plugin exposed via a HATEOAS REST API plus an MCP Python bridge for decompilation, analysis & binary manipulation"
readme = "README.md"
requires-python = ">=3.11"
authors = [
{name = "Ryan Malloy", email = "ryan@supported.systems"}
]
dependencies = [
"mcp>=1.22.0",
"requests>=2.32.3",
"fastmcp>=2.0.0",
]
[project.scripts]
ghydramcp = "bridge_mcp_hydra:main"
ghydramcp = "ghydramcp:main"
[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"
[tool.hatch.build.targets.wheel]
packages = ["."]
only-include = ["bridge_mcp_hydra.py"]
packages = ["src/ghydramcp"]
[tool.hatch.build]
sources = ["src"]
[tool.hatch.build.targets.sdist]
include = [
"/src",
"/README.md",
"/pyproject.toml",
]
[tool.ruff]
line-length = 100
target-version = "py311"
[tool.ruff.lint]
select = ["E", "F", "I", "W"]
ignore = ["E501"] # Line too long - handled by formatter

12
src/main/resources/Module.manifest
View File

@ -1,3 +1,9 @@
Manifest-Version: 1.0
GHIDRA_MODULE_NAME: GhydraMCP
GHIDRA_MODULE_DESC: A multi-headed REST interface for Ghidra for use with MCP agents.
# GhydraMCP Module Manifest
#
# This file lists third-party libraries bundled with this extension and their licenses.
# Module metadata (name, description, version) is defined in extension.properties.
#
# Format: MODULE FILE LICENSE: lib/filename.jar License Name
#
# Currently, GhydraMCP has no bundled third-party libraries.
# Gson is provided by Ghidra itself.