Windows 11 Is Collecting Data on Which Coding Agents and Developer Tools You Use
TL;DR Highlight
Reverse engineering shows Windows 11 logs which coding agents you use and reports it as optional diagnostics
Who Should Read
A developer who runs coding agents like Claude Code or Codex on Windows and worries about telemetry on their dev machine
Core Mechanics
- Agent names are hardcoded in MissionControlAggregator.dll
- The DLL compares against copilot.exe, claude.exe, codex.exe and agy.exe for CLI agents.
- Desktop slots check github.exe, claude.exe, codex.exe, cursor.exe and antigravity.exe.
- The feature is named CLICodingAgentUsage, feature ID 62580080.
- It also tracks general dev tools
- It checks node.exe, python.exe, dotnet.exe, gcc.exe, clang.exe, gdb.exe, javac.exe, devenv.exe, code.exe and more.
- Node.js, Python, .NET and Visual Studio slots also verify the publisher.
- The GCC/Clang compiler-debugger group has no publisher check.
- It listens to existing events instead of scanning
- It subscribes to KernelProcess.AppStateChangeSummary and Win32kTraceLogging.AppInteractivitySummary.
- For WSL it uses the imageName field of Lxss.Manager.Exec.
- So a regular Windows codex.exe is detected too, not only WSL.
- It records a bitmap, not usage time
- The Segments.SubCohorts report holds one bit per tool category.
- The author says runtime, focus time and resource usage from the source events appear to be discarded.
- The author found no prompt text or source code in the report.
- Renaming a file was enough to trigger it
- Copying cmd.exe to codex.exe set CLI slot 2, and the report value was 4 (1<<2).
- Copying it to gcc.exe set developer-tool slot 3, and the upload queue held SubCohorts [8].
- Uploaded when optional diagnostics are on
- The report is marked isRequired false, i.e. optional diagnostic data.
- In the author's US clean install of Windows 11 Home ARM64, the default was Include Optional with AllowTelemetry 3.
- The author says the evidence does not show the coding-agent feature is active on every installation.
Community Reactions
- Opinion that optional diagnostics are disclosed
- A commenter said Windows 11 offers Required and Optional toggles plus an app to browse event entries.
- They claimed turning optional data off has no effect on functionality or security.
- Opinion that telemetry for sandbox priorities is reasonable
- A commenter said Windows only launched its agent sandbox MXC a month ago (Oct 7) and Claude Code doesn't support it yet.
- They hope non-invasive telemetry helps Microsoft prioritize this.
- Experience of Defender quarantining Codex files
- A commenter said Windows Defender kept quarantining Codex's text config files so persistently they considered disabling it permanently.
- Question about Diagnostic Data Viewer
- A commenter asked whether this data shows up in Diagnostic Data Viewer.
- No answer appeared in the comments.
How to Apply
- Turn off optional diagnostic data
- Go to Settings → Privacy & security → Diagnostics & feedback now.
- Switch off Send optional diagnostic data and check every dev machine.
- Check the AllowTelemetry registry value
- Query AllowTelemetry from an admin PowerShell.
- If it is 3, add it to your team's PC checklist.
- Verify DiagTrack and AggregatorHost
- Run Get-Service DiagTrack and tasklist /m MissionControlAggregator.dll.
- Report whether the module is loaded against your org policy.
- Decide a policy on agent-usage exposure
- Agree with your security team whether agent usage may leave the company.
- If not, enforce disabling optional diagnostics via group policy.
Try It
# Paste into an admin PowerShell (based on the article's check commands)
Get-Service DiagTrack | Format-Table Name,Status,StartType -AutoSize
Get-Process AggregatorHost | Format-Table ProcessName,Id -AutoSize
tasklist /m MissionControlAggregator.dll
# Check the diagnostic data setting (3 means Include Optional)
Get-ItemProperty 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\DataCollection' -Name AllowTelemetry -ErrorAction SilentlyContinue
# Next: turn off Send optional diagnostic data in Settings → Privacy & security → Diagnostics & feedback and re-checkTerminology
Related Papers
LLM assisted reverse engineering
Pointing an agent at binaries with skills and MD notes made years-long hacks feasible
Metrics-in-the-Loop: Complexity Feedback Reshapes Agent-Written Code
Injecting per-turn complexity reports into coding agents cut code erosion without changing correctness
Why is Gemini 4 Carbon taking so long?
Carbon took nearly an hour on bug fixes that Astra and Opus finished in minutes, mostly from extra unrequested work.
How the Claude Code team thinks about verification
E2E, property and differential tests plus verification rules let Claude run in parallel.
Running coding agents without handing them my credentials
Credentials stay on the laptop; agents run commands there via Touch ID approval
The Silent Cost of Conversation History in LLM Reasoning
Keeping full conversation history dropped accuracy from 0.6 to 0.1 within 20 turns; a verified state summary prevented it.
Anthropic upgrades Claude Code Projects to run parallel AI workers
Claude Code Projects now has a coordinator that splits work into parallel threads and opens PRs.
Show HN: AgentRun – open-source, Self-improving Managed Agents
Agents save notes and skills after each run, cutting code review cost to 18 cents