Context7
Feeds AI the latest docs for libraries and frameworks before it answers, so code it writes does not use outdated, removed functions.
Description from the maker (original English)
Windows x64dbg/x32dbg debugging with guarded sessions, breakpoints, memory and runtime evidence.
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Download the maker’s .mcpb file
The link goes to the maker’s own file; we do not host or repackage it. Open the file and Claude Desktop installs it.
Run this in your terminal
Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 8, 2026
The registry name was verified with the GitHub account rison1337, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.1.9) reached the registry 2 months ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need MCPB on your computer.
In the same category: Context7, Codebase Memory, Exa
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Feeds AI the latest docs for libraries and frameworks before it answers, so code it writes does not use outdated, removed functions.
Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
https://github.com/user-attachments/assets/07b813eb-4175-4f14-b21b-602548906398
x64dbg MCP gives an MCP client guarded control of a real x64dbg or x32dbg session. It turns live debugging into a replayable reverse-engineering workflow: bind the exact process, observe runtime facts, recover useful artifacts, validate them independently, and hand the evidence to IDA.
| Workflow | Representative tools | Result |
|---|---|---|
| Launch and bind | InitDebuggee, AttachToProcess, LaunchFileUnderDebugger, WaitForBreakpoint | Reproducible process and session identity |
| Live control | RegisterGet, RegisterSet, MemoryRead, MemoryWrite, DebugSetBreakpoint, SetHardwareBreakpoint | Controlled execution, memory and breakpoint changes |
| Runtime evidence | RunNativeTrace, GetNativeTrace, StartApiTrace, StartHeapTrace, GetBasicBlockCoverage, WaitForBreakpointCapture | Instruction paths, API/heap calls, exceptions and executed blocks |
| Key and unpack recovery | SearchStrings, ScanMemoryStrings, PatternFindMem, FindOEP, RunUntilOEP, FindIATCandidates, InspectRuntimeIAT | Comparisons, strings, OEP and runtime import candidates |
| Dump and repair | WriteMiniDump, DumpModuleRaw, DumpPeFromMemory, ScanMemoryForPEImages, FixDumpImports, ValidateDump, ExportPatchedFile | Replayable dumps and independently checked PE artifacts |
| IDA evidence handoff | ExportRuntimeEvidence, ImportStaticAnnotations, SyncBreakpoints, ResolveModuleRva | Hash/RVA-addressed comments, labels, coverage and API facts |
The complete catalog, parameters and response contracts are in the tool reference. The compact profile keeps routine model-facing responses short; detail="full" and the full profile expose the complete evidence when it is needed.
Download the combined Windows bundle from Releases. It contains both native plugins and the Python backend:
plugins\MCPx64dbg.dp64 for x64dbgplugins\MCPx64dbg.dp32 for x32dbgruntime\src and runtime\requirements.txtruntime\tools\bin\managed_probeClose x64dbg/x32dbg. Open PowerShell in the folder containing the downloaded ZIP, then run the block below. It asks where to keep the MCP bundle and where x64dbg is already installed; no drive or installation directory is assumed. Keep the same PowerShell window open for the client-specific commands below.
$Archive = Get-ChildItem -File .\x64dbg-mcp-windows-*.zip |
Sort-Object LastWriteTime -Descending | Select-Object -First 1 if (-not $Archive) { throw 'The x64dbg MCP release ZIP was not found in this folder.' }
$BundleRoot = Read-Host 'Absolute folder where x64dbg MCP should be extracted' $X64dbgRoot = Read-Host 'Absolute folder containing the x64 and x32 x64dbg folders' $BundleRoot = [IO.Path]::GetFullPath(
$X64dbgRoot = (Resolve-Path -LiteralPath (
if (-not (Test-Path -LiteralPath (Join-Path $X64dbgRoot 'x64\x64dbg.exe'))) { throw "x64dbg.exe was not found under $X64dbgRoot\x64" } if (-not (Test-Path -LiteralPath (Join-Path $X64dbgRoot 'x32\x32dbg.exe'))) { throw "x32dbg.exe was not found under $X64dbgRoot\x32" }
Expand-Archive -LiteralPath $Archive.FullName -DestinationPath $BundleRoot -Force Copy-Item (Join-Path $BundleRoot 'plugins\MCPx64dbg.dp64') (Join-Path $X64dbgRoot 'x64\plugins\MCPx64dbg.dp64') -Force Copy-Item (Join-Path $BundleRoot 'plugins\MCPx64dbg.dp32') (Join-Path $X64dbgRoot 'x32\plugins\MCPx64dbg.dp32') -Force
$RuntimeRoot = Join-Path $BundleRoot 'runtime' Set-Location $RuntimeRoot py -3.11 -m venv .venv .\.venv\Scripts\python.exe -m pip install -r requirements.txt
$PythonExe = (Resolve-Path .\.venv\Scripts\python.exe).Path $Launcher = (Resolve-Path .\src\mcp_stdio_launcher.py).Path [pscustomobject]@{ PythonExe = $PythonExe Launcher = $Launcher X64dbgRoot = $X64dbgRoot }
Add the server to ~\.codex\config.toml. Replace the three angle-bracketed values with the absolute paths printed by the setup block. TOML single-quoted strings preserve Windows backslashes as written.
[mcp_servers.x64dbg]
command = ' ' args = [' '] startup_timeout_sec = 90
[mcp_servers.x64dbg.env] X64DBG_ROOT = ' ' X64DBG_MCP_TOOL_PROFILE = 'compact'
For normal target startup, call InitDebuggee directly with the EXE path. It detects x86/x64, starts the matching debugger from X64DBG_ROOT, waits for the bridge and opens the target. A separate BridgeHello preflight or manual debugger-path search is not required.
Optional installation check: start x64dbg or x32dbg and verify the bridge:
& $PythonExe (Join-Path $RuntimeRoot 'src\x64dbg.py') GetDebuggerPluginStatus
GetScyllaHideStatus.installed and integrationReady describe the MCP InjectorCLI/HookLibrary backend. guiPluginPresent describes only the optional x64dbg GUI plugin; it is not required for MCP injection.
ScyllaHide is opt-in: launch tools default to use_scyllahide="off", and auto also skips injection. Use force with an explicit profile when needed. The CLI backend stages a private INI beside the injector and targets the exact PID. Its compatibility configuration disables NtContinueHook and KillAntiAttach together to avoid the reproduced continuation crash; the result reports these overrides. This does not guarantee that every third-party profile is compatible with every target or Windows build.
Add the same stdio server to Claude Code with the user scope. This command uses the paths selected by the setup block instead of embedding a machine-specific location:
$ClaudeServer = [ordered]@{
command = $PythonExe args = @($Launcher) env = [ordered]@{ X64DBG_ROOT = $X64dbgRoot X64DBG_MCP_TOOL_PROFILE = 'compact' } } | ConvertTo-Json -Depth 4 -Compress claude mcp add-json x64dbg $ClaudeServer --scope user claude mcp list
Alternatively, run claude mcp add and enter the same command, arguments and environment interactively.
Use the client's MCP JSON configuration. For Claude Desktop on Windows, the file is %APPDATA%\Claude\claude_desktop_config.json. Generate a JSON block with the actual paths selected above:
$ClientConfig = [ordered]@{
mcpServers = [ordered]@{ x64dbg = [ordered]@{ command = $PythonExe args = @($Launcher) env = [ordered]@{ X64DBG_ROOT = $X64dbgRoot X64DBG_MCP_TOOL_PROFILE = 'compact' } } } } $ClientConfig | ConvertTo-Json -Depth 6
Cursor, VS Code MCP, Windsurf and other stdio clients use the same command/args/env contract; only the location of their JSON file differs.
git clone https://github.com/rison1337/x64dbgMCP.git
Set-Location x64dbgMCP cmake -S . -B build -G "Visual Studio 17 2022" -DX64DBG_DOWNLOAD_SDK=ON cmake --build build --target all_plugins --config Release
$X64dbgRoot = (Resolve-Path -LiteralPath (Read-Host 'x64dbg installation folder')).Path Copy-Item build\build64\Release\MCPx64dbg.dp64 (Join-Path $X64dbgRoot 'x64\plugins\MCPx64dbg.dp64') -Force Copy-Item build\build32\Release\MCPx64dbg.dp32 (Join-Path $X64dbgRoot 'x32\plugins\MCPx64dbg.dp32') -Force
The IDA workflow is designed for the rison1337/ida-pro-mcp-fusion fork. x64dbg MCP does not pretend to be an IDA replacement and does not silently mutate an unrelated database. Instead:
ExportRuntimeEvidence writes a versioned evidence document containing the target SHA-256, architecture, module RVAs, executed blocks, API calls, comments, labels and functions.ResolveModuleRva and the coordinator normalize live addresses to the static image identity.tools/ida_evidence_coordinator.py validates that the open Fusion database has the same SHA-256 and architecture, then creates deterministic set_name, set_comments, define_func and coverage/API comment actions.This is an explicit artifact/protocol handoff, so it remains inspectable and retryable. See the coordinator and the Fusion README for the IDA-side worker and cache model.
CI covers the Python contract and both native plugin architectures. The release workflow also checks the packaged stdio server and managed probes. Live debugger gates run locally and retain JSON reports; see Contributing. Analyze untrusted binaries inside a disposable VM.
GPL-3.0. Based on x64dbgMCP by Sam W (Wasdubya). See NOTICE for attribution.
x64dbg MCP даёт MCP-клиенту защищённое управление настоящей сессией x64dbg или x32dbg. Он превращает живую отладку в воспроизводимый reverse-engineering workflow: привязать точный процесс, снять runtime-факты, восстановить полезные артефакты, независимо проверить результат и передать факты в IDA.
| Этап | Примеры MCP tools | Что получается |
|---|---|---|
| Запуск и привязка | InitDebuggee, AttachToProcess, LaunchFileUnderDebugger, WaitForBreakpoint | Воспроизводимый процесс и точная session identity |
| Управление | RegisterGet, RegisterSet, MemoryRead, MemoryWrite, DebugSetBreakpoint, `SetHardwareBr
Only the beginning is shown; read the rest in the repo