Gitlab (zereight) MCP Server
@zereight/mcp-gitlab
npm
v2.1.42
Source verified
Published by zereight — publish provenance cryptographically ties this package to that repository. That is proof of origin, not an official vendor package.
Started at 100 and nothing was deducted: the deterministic scan raised no scored threat in the published surface it inspected. Capability observations and advisory notes are recorded but never lower the score.
Findings 16
high Shell/command execution in server code (build/oauth.js)MTC-SRC-002
In the server's implementation (`build/oauth.js:8`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/oauth.js
high Shell/command execution in server code (build/test/config-allowed-groups.test.js)MTC-SRC-002
In the server's implementation (`build/test/config-allowed-groups.test.js:15`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/config-allowed-groups.test.js
high Shell/command execution in server code (build/test/sse-auth-guard.test.js)MTC-SRC-002
In the server's implementation (`build/test/sse-auth-guard.test.js:2`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/sse-auth-guard.test.js
high Shell/command execution in server code (build/test/stateless/config-ttl.test.js)MTC-SRC-002
In the server's implementation (`build/test/stateless/config-ttl.test.js:21`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/stateless/config-ttl.test.js
high Shell/command execution in server code (build/test/streamable-http-dns-rebinding.test.js)MTC-SRC-002
In the server's implementation (`build/test/streamable-http-dns-rebinding.test.js:2`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/streamable-http-dns-rebinding.test.js
high Shell/command execution in server code (build/test/streamable-http-static-token-auth.test.js)MTC-SRC-002
In the server's implementation (`build/test/streamable-http-static-token-auth.test.js:2`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/streamable-http-static-token-auth.test.js
high Shell/command execution in server code (build/test/test-ci-catalog.js)MTC-SRC-002
In the server's implementation (`build/test/test-ci-catalog.js:3`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-ci-catalog.js
high Shell/command execution in server code (build/test/test-ci-lint.js)MTC-SRC-002
In the server's implementation (`build/test/test-ci-lint.js:3`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-ci-lint.js
high Shell/command execution in server code (build/test/test-ci-variables.js)MTC-SRC-002
In the server's implementation (`build/test/test-ci-variables.js:3`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-ci-variables.js
high Shell/command execution in server code (build/test/test-create-repository.js)MTC-SRC-002
In the server's implementation (`build/test/test-create-repository.js:3`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: child_process
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-create-repository.js
medium Hardcoded egress to an external endpoint (build/test/sse-auth-guard.test.js)MTC-SRC-003
In the server's implementation (`build/test/sse-auth-guard.test.js:59`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${LOOPBACK}:${port}/health`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/sse-auth-guard.test.js
medium Hardcoded egress to an external endpoint (build/test/streamable-http-dns-rebinding.test.js)MTC-SRC-003
In the server's implementation (`build/test/streamable-http-dns-rebinding.test.js:61`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${hostWithPort(HOST, port)}/health`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/streamable-http-dns-rebinding.test.js
medium Hardcoded egress to an external endpoint (build/test/streamable-http-static-token-auth.test.js)MTC-SRC-003
In the server's implementation (`build/test/streamable-http-static-token-auth.test.js:60`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${HOST}:${port}/health`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/streamable-http-static-token-auth.test.js
medium Hardcoded egress to an external endpoint (build/test/test-all-transport-server.js)MTC-SRC-003
In the server's implementation (`build/test/test-all-transport-server.js:220`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${HOST}:${port}/mcp`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-all-transport-server.js
medium Hardcoded egress to an external endpoint (build/test/test-remote-downloads.js)MTC-SRC-003
In the server's implementation (`build/test/test-remote-downloads.js:120`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${HOST}:${serverPort}/downloads/job-artifacts?project_id=${TEST_PROJECT_ID}&job_id=${TEST_JOB_ID}`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/test-remote-downloads.js
medium Hardcoded egress to an external endpoint (build/test/utils/server-launcher.js)MTC-SRC-003
In the server's implementation (`build/test/utils/server-launcher.js:205`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch(`http://${HOST}:${port}/health`
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server build/test/utils/server-launcher.js
What this scan could not see
- No tools were enumerated, so prompt-injection, capability and toxic-flow analysis had no tool surface to inspect. To grade a package’s real runtime tools, scan the running server: --command "npx -y <package>".
Versions
Scan history per published version. The engine is deterministic — the same version always yields the same score, so a changed score means the package itself changed.
| Version | Score | Findings | Engine | Scanned |
|---|---|---|---|---|
v2.1.42 latest |
A 100/100 | 16 | 1.5.0 | 2026-07-22 |
Embed this score
Show this server's live Trust Score in your README, docs or website. The badge is served straight from the registry and updates automatically after every rescan — no API key needed. It links back to this page, so anyone who sees the grade can also read the findings behind it instead of taking a number on faith.
Verify this score yourself
The score above is reproducible: the same package version always yields the same result. Run it locally or over the free API — no account, no LLM, fully deterministic.
npx mcptrustchecker scan @zereight/mcp-gitlab --onlineOther implementations of Gitlab 22
Independent packages implementing the same tool, scanned with the same engine. Compare all 23 side by side →
Advanced server for working with GitLab.
Server providing access to GitLab.
Security scan results for the Gitlab (yoda.digital) MCP server.
Security scan results for the Gitlab (ttpears) MCP server.
Manages GitLab epics, issues, and milestones.
Security scan results for the Gitlab (rokrokss) MCP server.
More in AI & Agents
Security scan results for the 37soul MCP server.
Security scan results for the 402 MCP server.
Security scan results for the A2a Governance Bridge MCP server.
Security scan results for the A2anet MCP server.
Security scan results for the Abis MCP server.
Security scan results for the Acr MCP server.