winrdp-mcp@0.1.11
CAUTION· not startedPublisher claims: "Zero-config MCP server to provision and fully administer any Windows RDP box (Win10/11/Server) — built for Claude & Claude Code." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Not started: dependency install failed; the server itself was not observed.- Tested
- the dependency install step · every file read, file written and host contacted by what ran
- Not tested
- server startup · the tool list · tool behavior
What this means for the label: Caution here covers what ran. The server's tools were not exercised with a real credential, so nothing about their behavior is asserted; the credential summary below says what to do with that.
Before you give it your token
- It will ask for
WINRDP_DEFAULT_WINRM_AUTHConfiguration, not credentials: WINRDP_AUDIT_LOG, WINRDP_CONFIRM_DESTRUCTIVE, WINRDP_DEBUG, WINRDP_HOME, WINRDP_PROFILE, WINRDP_VAULT_KEY, WINRDP_WINRM_OP_TIMEOUT- Where it talks to
- no vendor host named in the codeAlso sysinternals.com, nssm.cc, python.org, chocolatey.org, aka.ms — not the publisher's domain; check what for before connecting.
- What we saw when it ran
- no credential files read · no network
- What we could not test
- server startup · the tool list · tool behavior
Connect it with a scoped credential, under the guard, and read the third-party hosts first. The code names sysinternals.com, nssm.cc, python.org, chocolatey.org, aka.ms as well as its own vendor. The guard injects WINRDP_DEFAULT_WINRM_AUTH only into requests to the host you bind (and its subdomains) and refuses the rest.
Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; not started: dependency install failed; the server itself was not observed) on 2026-11-21 and scored it 75/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Privileged sudo command; README links to an external installer; Privileged sudo command. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 0 s · 2 scenarios · 0 deviationsStatic findings
proc = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.sr = transport.run_sh("sudo -n true && echo WINRDP_SUDO_OK", timeout=30)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.https://example.com/tool.exeFix: Pin versions and sources; vendor anything fetched at install time.subprocess.run(["cmdkey", f"/generic:TERMSRV/{h.host}",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.proc = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."" if self.sudo else "sudo: a password is required", 0 if self.sudo else 1)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.host, `elevated=True` prefixes `sudo -n`; `as_user`/`kill_on_timeout` are Windows-only.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.subprocess.run(["cmdkey", f"/generic:TERMSRV/{h.host}", f"/user:{user}",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.subprocess.Popen(["mstsc", f"/v:{h.host}:{h.rdp_port}"])Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.runner = "sudo -n " + runnerFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.proc = subprocess.run(["cmd.exe", "/c", command], capture_output=True, timeout=timeout)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.subprocess.run("clip", input=line.encode("utf-16-le"), check=True,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.if "sudo -n true" in script:Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.proc = subprocess.run(shlex.split(command), capture_output=True, timeout=timeout)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.cp = subprocess.run(["cmdkey", f"/generic:TERMSRV/{h.host}",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.The script sleeps well past the deadline and only then writes `marker`. If the killFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.subprocess.Popen(["mstsc", f"/v:{h.host}:{h.rdp_port}"])Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.return ExecResult("WINRDP_SUDO_OK\n" if self.sudo else "",Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.r = subprocess.run(["icacls", str(path), "/inheritance:r", "/grant:r",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.f"root={uid == '0'}, sudo -n={'ok' if sudo_ok else 'unavailable'}"))Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.self.sudo = sudoFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.`kill_on_timeout`: kill the remote process when `timeout` expires instead ofFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# process whose command line references that script (the task's own tree) and kill it.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell._log.warning("kill_on_timeout: could not kill marker %s on %s: %s",Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.f"timeout kill it instead of leaving it running.")Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.we meant to kill has already run to completion by the time we look for it. AFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# kill processes spawned from the task's staged script (and their children, /T)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.except Exception as e: # noqa: BLE001 — the kill is best-effortFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""Stop a running scheduled task by name and kill its process tree — the clean way toFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.timer.join(60) # let the kill finish so the pid list is accurateFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# the caller didn't ask us to kill — unregistering a task that is still runningFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# call, and it needs room to kill the tree before the transport gives up.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""Stop a scheduled task AND kill the process tree it spawned; return the killed pids.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.`kill_on_timeout`: kill the remote python process when `timeout` expires — byFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.V.package("pkg; rm -rf /")Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.V.package("pkg; rm -rf /")Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# Destructive tools: delete data, kill processes, cut access, or reboot. Clients shouldFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.- [System](#system) — `winrdp_mcp/tools/system.py`Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Cost to the agent
Tokens estimated at four characters each. The tool list is re-read by the agent on every turn, so it is the number to compare between servers that do the same job: ranked across the index.
Known vulnerabilities
Not checked against OSV yet; the weekly check has not reached this version.
What would make this Verified
5 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Unclaimed publisher≈ +8 on the totalClaim emog33k: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Privileged sudo command≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- README links to an external installer≈ +4 on the totalPin versions and sources; vendor anything fetched at install time.
- Python subprocess execution detected≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
- Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- Destructive file system command detected≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.