Graph (graph-tl) MCP Server

@graph-tl/graph npm v0.2.20

Published by @graph-tl — no publish provenance, so origin is unverified, but the source is public: the repository link below is self-declared yet readable, so you can inspect the code before adopting it.

Agent-native persistent task graph. MCP server for agent planning and execution across sessions.

Trust grade
B
82/100
Last scanned get badge →
Trust
B · 82/100
Adoption risk for you: the threat score, then adjusted down for blast radius, publisher verification and how much the scan could see. Deterministic; every point is auditable.
Capability
High
Blast radius if it went rogue — what the server’s tools could reach. Independent of trust.
Coverage
Source
How much the scan could actually inspect. Shallow coverage is stated, never hidden.
Share this Trust Score
𝕏 Share LinkedIn Reddit
B Why this grade threat 89 − adoption risk = 82/100

The grade answers one question — how safe is this server for you to adopt — so it is computed in two auditable stages. Nothing below is an opinion or an LLM's guess; every line is a real term the deterministic engine applied, and the same input always yields the same number.

1. Threat score — 100 − 11 = 89. What the published surface and source actually contain:

PointsWhat was foundCategory
−11 Untrusted input concatenated into a command sink ×3 MTC-SRC-009 injection

2. Client adoption risk — 89 − 7 = 82. Three small, subtract-only factors that reflect your risk in adopting it — a clean scan proves less on a powerful, unverified or barely-inspectable package, so the grade says so plainly:

PointsAdoption-risk factor
−6 capability blast radius (high) — client exposure if the model is manipulated
−1 publisher verification (public source) — no provenance, but the source is public and inspectable

Capability observations and info notes are shown under Findings but never scored. Open any row's finding below for the file, line and evidence behind a deduction.

Findings 7

high Shell/command execution in server code (dist/index.js)MTC-SRC-002

In the server's implementation (`dist/index.js:47`): 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: cSync } = await import("child_process"); console.log("Clearing npx cache..."); try { execSync("npx clear-npx-cac

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 dist/index.js

high Shell/command execution in server code (dist/server-BU5MZNHH.js)MTC-SRC-002

In the server's implementation (`dist/server-BU5MZNHH.js:2665`): 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: port { execSync } from "child_process"; function tryGitCommits(since, max = 10) { try { const sinceArg = since ? `

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 dist/server-BU5MZNHH.js

high Shell/command execution in server code (dist/ui-6UQ75H4I.js)MTC-SRC-002

In the server's implementation (`dist/ui-6UQ75H4I.js:5`): 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: ; import { exec } from "child_process"; import { existsSync, readFileSync, readdirSync } from "fs"; import { dirname, jo

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 dist/ui-6UQ75H4I.js

medium Hardcoded egress to an external endpoint (dist/server-BU5MZNHH.js)MTC-SRC-003

In the server's implementation (`dist/server-BU5MZNHH.js:2997`): 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: { const res = await fetch(`https://registry.npmjs.org/${PKG_NAME}/latest`); if (!res.ok) return; const data

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 dist/server-BU5MZNHH.js

medium Untrusted input concatenated into a command sink (dist/index.js)MTC-SRC-009

In the server's implementation (`dist/index.js:128`): A shell/process command assembled from concatenated or interpolated values is command injection when any part is attacker-influenced — the OWASP canonical RCE flow. Verify what reaches the interpolated value. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: }).trim(); notes = execSync(`git log --oneline ${lastTag}..HEAD~1`, { encoding: "utf-8", stdio: ["pipe", "pipe", "p

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 dist/index.js

medium Untrusted input concatenated into a command sink (dist/server-BU5MZNHH.js)MTC-SRC-009

In the server's implementation (`dist/server-BU5MZNHH.js:2669`): A shell/process command assembled from concatenated or interpolated values is command injection when any part is attacker-influenced — the OWASP canonical RCE flow. Verify what reaches the interpolated value. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: ""; const output = execSync( `git log ${sinceArg} --oneline --no-decorate -${max}`, { encoding: "utf-8"

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 dist/server-BU5MZNHH.js

medium Untrusted input concatenated into a command sink (dist/ui-6UQ75H4I.js)MTC-SRC-009

In the server's implementation (`dist/ui-6UQ75H4I.js:2549`): A shell/process command assembled from concatenated or interpolated values is command injection when any part is attacker-influenced — the OWASP canonical RCE flow. Verify what reaches the interpolated value. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: "start" : "xdg-open"; exec(`${cmd} ${url}`); } function discoverDbs() { if (process.env.GRAPH_DB) { return exist

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 dist/ui-6UQ75H4I.js

Tools 25

Each tool and what it can reach — statically extracted from the published source.

  • graph_agent_configno sensitive capability
  • graph_connectno sensitive capability
  • graph_contextno sensitive capability
  • graph_historyno sensitive capability
  • graph_knowledge_auditno sensitive capability
  • graph_knowledge_deleteno sensitive capability
  • graph_knowledge_readno sensitive capability
  • graph_knowledge_searchno sensitive capability
  • graph_knowledge_writeno sensitive capability
  • graph_knowledge_write_batchno sensitive capability
Show 15 more tools ↓
  • graph_nextno sensitive capability
  • graph_onboardno sensitive capability
  • graph_openno sensitive capability
  • graph_planno sensitive capability
  • graph_queryno sensitive capability
  • graph_resolveno sensitive capability
  • graph_restructureno sensitive capability
  • graph_retrono sensitive capability
  • graph_roadmapno sensitive capability
  • graph_statusno sensitive capability
  • graph_treeno sensitive capability
  • graph_updateno sensitive capability
  • Knowledge Entryno sensitive capability
  • Project Knowledgeno sensitive capability
  • Project Treeno sensitive capability

What this scan could not see

Versions 1

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.

VersionScoreFindingsEngineScanned
v0.2.20 latest B 82/100 7 1.13.0 2026-09-07

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.

MCP Trust Score: B · 82/100
Markdown (GitHub README)
[![MCP Trust Score](https://mcptrustchecker.com/registry/graph-tl-graph/badge.svg)](https://mcptrustchecker.com/registry/graph-tl-graph)
HTML
<a href="https://mcptrustchecker.com/registry/graph-tl-graph"><img src="https://mcptrustchecker.com/registry/graph-tl-graph/badge.svg" alt="MCP Trust Score" height="20"></a>
Prefer shields.io styling? Point it at https://mcptrustchecker.com/registry/graph-tl-graph/badge.json via https://img.shields.io/endpoint?url=…

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 @graph-tl/graph --online

Use the free API → How scoring works

Other implementations of Graph 4

Independent packages implementing the same tool, scanned with the same engine. Compare all 5 side by side →

More in AI, Memory & Reasoning