vestibule-mcp@0.1.2
CAUTION· not startedPublisher claims: "A local, kernel-isolated code-execution sandbox for AI agents, exposed as an MCP server." 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
- no credential variables read in the code
- Where it talks to
- no vendor host named in the codeAlso docker.com — 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 under the guard. Not started: dependency install failed; the server itself was not observed. The code names no credential variable; the guard pins the network to what the honeypot saw and logs every call.
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 2026-11-20 and scored it 72/100 (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; References a credential file in the home directory; Image alt text carrying instructions. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 0 s · 2 scenarios · 0 deviationsStatic findings
`vestibule-warden -S -P -c "exec(open(<fd>,'rb').read())"`, and this file is that programFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."vestibule-warden", "-S", "-P", "-c", f"exec(open({source},'rb').read())",Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.# Outlasts the finisher's kill (5 s) + rm -f (5 s), with slack.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."(`sudo usermod -aG docker $USER`), log out and back in, then restart your agent. "Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.r = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.proc = subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.log.warning("run %s: timeout after %ds; container kill detached", run_id, timeout_s)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.# overflow kill the container early rather than letting it chatter on.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.os.execve(f"{directory}/{name}", cfg["argv"], env)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.including on guest timeout, where the kill is fully detached (§4.5, amended by theFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.~/.docker/config.json puts those two words in a WARNING on the first line instead. It doesn'tFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.and a POSIX parent's its owner's pid, so a sweep can remove what a hard kill left behindFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Terminal recording. A short program runs twice. On the machine itself it reads a fake key file and connects to 1.1.1.1. Run through Vestibule, the same program Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Terminal recording. A short program runs twice. On the machine itself it reads a fake key file and connects to 1.1.1.1. Run through Vestibule, the same program Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.start_new_session=True, # out of reach of a PID kill and of the SDK's killpgFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# POSIX: its own process group, so a kill reaches the guest's childrenFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# S7-D1 shutdown budget. MCP clients close stdin, wait ~2 s, then kill the server'sFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# processes that would outlive the server's exit: kill them with the CLIs.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.and the set-id/sticky bits but keep owner rwx, so the cache can be deleted without a chmodFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.except Exception as e: # never let an exception kill the sessionFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.if truncated: # as M1's CLI reports after its killFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.runs in its own session, so neither a kill of the server's PID nor a kill of its processFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# normal teardown used to kill it, but POSIX now exits with os._exit (M2 step 1b).Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.~/.docker/config.json puts those two words in a WARNING on the first line instead. It doesn'tFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.# Guest processes still running, so the POSIX shutdown hand-off can kill themFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""Snapshot, kill CLIs, start the cleaner, exit 0. Synchronous: no await anywhere, soFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""S4-D8, from the helper's line of guest/'s counters: `oom_kill` counts any OOM kill inFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# used to kill them.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.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
8 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Unclaimed publisher≈ +8 on the totalClaim victorperson999: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Hidden instructions≈ +6 on the totalRemove HTML comments, zero-width characters and white-on-white text that carry instructions. Everything the agent is meant to read should be visible to the person reading the file.
- Privileged sudo command≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- References a credential file in the home directory≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
- Python exec or eval of computed code≈ +2 on the totalPut hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
- Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.