rutherford-mcp-server@3.3.0
CAUTION· install onlyPublisher claims: "An MCP server that orchestrates a crew of agentic coding agents over the Agent Client Protocol (ACP)." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Nothing executable in the bundle ran; the trace covers the install step only.- Tested
- every file read, file written and host contacted by what ran
- Not tested
- use of the skill by an agent
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 agentclientprotocol.com — not the publisher's domain; check what for before connecting.
- What the honeypot saw
- no credential files read · no network
- What we could not test
- use of the skill by an agent
Connect it under the guard. Nothing executable in the bundle ran; the trace covers the install step only. 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 (1 scenario; nothing executable in the bundle ran; the trace covers the install step only) on 2026-11-16 and scored it 71/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Access to sensitive file paths; Access to sensitive file paths; Access to sensitive file paths. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 0 s · 1 scenarios · 0 deviationsStatic findings
"""Wait for the command to finish (bounded by ``timeout_s``); kill + reap and report -1 on a timeout."""Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.so :func:`bedrock_remediation_hint` instead points the user at the per-agent ``[agents.<id>.env]`` config thatFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.f"[agents.{agent}.env] in config, then re-run doctor agent={agent}"Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.acp``), an npm-installed one with the npx args (``gemini --acp``).Fix: Pin versions and sources; vendor anything fetched at install time.#: Bedrock/Vertex (or enterprise-wrapped) provider rejected, pointing at the per-agent ``[agents.<id>.env]``Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.# over is the same shape the subprocess got -- os.environ with the descriptor's [agents.<id>.env] overridesFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.completed = subprocess.run( # noqa: S603 - fixed `git` argv0 plus internal subcommands, no shellFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."""Await one shared kill task so cancellation of a caller cannot strand the process or lose ownership."""Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.stripped = text.lstrip("") # tolerate a UTF-8 BOM on a Windows-authored fileFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.kill / release. ``read_only`` / ``propose`` keep denying terminal execution.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.#: kill the reader that exists to diagnose it.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# Any userinfo: `https://user:secret@host/`, or a token as the username (the git / PAT form). A top-levelFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.variables go in ``[agents.<id>.env]`` or the server's environment.Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.# seat gets a targeted remediation hint (the per-agent [agents.<id>.env] fix). Signature- and host-gated, soFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.process = subprocess.Popen( # noqa: S603, ASYNC220 - long-lived process drained off-threadFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.#: Maximum wait after a hard kill before terminal teardown proceeds to descendant reap and reader cancellation.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.#: The pre-kill descendant enumeration. Bounded because it runs before the adapter kill: a stage that canFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.completed = subprocess.run( # noqa: S603 - argv is `npm` plus a curated constant, never caller inputFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.f" [agents.{descriptor.id}.env]\n"Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.# an owner-only config that can hold [agents.*.env] values.Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.# fastmcp.settings.transport, a pydantic-settings field with env_prefix "FASTMCP_" and .env support,Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.#: Deadline on the pre-kill descendant enumeration. Mirrors the session teardown bound: losing theFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.inference profile). The hint describes the per-agent ``[agents.<id>.env]`` fix -- which lives outside theFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose."""The ``(bin_name, args)`` candidates from the npx distribution (the likely installed-bin names)."""Fix: Pin versions and sources; vendor anything fetched at install time.* It must be bounded. An unbounded await never returns, so the caller never reaches its kill and itsFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Shared by every pre-kill snapshot on purpose. Each caller runs this BEFORE hard-killing the process itFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.line would kill the reader whose only job is to explain such an agent.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""Terminate ``procs``, wait briefly, then kill any survivor. Best-effort; never raises."""Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# * The descendant snapshot is complete before this call, so a hard kill freezes the tree before reap.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# its own budget while the kill and reap carry on. Before that, Windows refuses to delete a liveFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."""Hard-kill the direct ACP adapter and bound the wait for its process handle to signal exit."""Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.completed = subprocess.run( # noqa: S603 - fixed `git` argv0; `args` are internal, `working_dir` is vettedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.Cost to the agent
Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.
Known vulnerabilities
Not checked against OSV yet; the weekly check has not reached this version.
What would make this Verified
9 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Unclaimed publisher≈ +8 on the totalClaim chapmanjw: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Access to sensitive file paths≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
- Hidden zero-width characters detected≈ +4 on the totalKeep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
- URL with embedded credentials≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
- Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- npx invocation of non-well-known package≈ +2 on the totalPin versions and sources; vendor anything fetched at install time.