registry / @withfudge/design / 2.0.7
@withfudge/design@2.0.7
Installer and agent guidance for the hosted Fudge design-corpus MCP server.
AI Security Review
scanned 17d ago · by lpm-firewall-aiSource inspection found no install lifecycle hook or confirmed attack behavior. The flagged process and network calls belong to user-invoked CLI features and are scoped to browser login, Keychain access, and Fudge workspace setup.
Static reason
One or more suspicious static signals were detected.
Trigger
The user invokes the login command or the workspace connect/add command.
Impact
No confirmed unauthorized credential transfer, install-time mutation, or other attack impact was established.
Mechanism
Login exchanges a temporary code with the configured Fudge API. Workspace connection uses the authenticated session with that API and saves scoped npm configuration.
AI rationale
Source inspection found no install lifecycle hook or confirmed attack behavior. The flagged process and network calls belong to user-invoked CLI features and are scoped to browser login, Keychain access, and Fudge workspace setup.
Evidence
project .npmrcuser .npmrcFudge session storage
Network endpoints2
api.withfudge.com127.0.0.1
Decision evidence
public snapshotAI called this Clean at 96.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- package.json exposes a CLI and has no install lifecycle hook; its scripts are build, prepack, test, and typecheck.
- Login is an explicit CLI action; it exchanges a temporary login code with the configured API, whose default is api.withfudge.com, then stores the returned session.
- The flagged child-process calls support browser launch and macOS Keychain operations; source shows no shell command used to harvest or transfer credentials.
- Workspace connection is explicitly invoked and sends the saved session to the configured Fudge API; it writes the registry token to the user npm config and only the @fudge registry setting to the project config.
Evidence against
- The install and network signals resolve to CLI features invoked by the user, with no confirmed attack behavior or automatic install-time action.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetwork
HighEntropyStrings
Source & flagged code
4 flaggeddist/ops/auth.jsView file
4import { join } from "node:path";
L5: import { execFile } from "node:child_process";
L6: import { promisify } from "node:util";
High
dist/ops/login.jsView file
107const args = process.platform === "win32" ? ["/c", "start", "", url] : [url];
L108: const child = spawn(command, args, { stdio: "ignore", detached: true });
L109: child.unref();
...
L118: function apiBase(value) {
L119: return (value || process.env.FUDGE_API_BASE || "https://api.withfudge.com").replace(/\/+$/, "");
L120: }
High
Same File Env Network Execution
A single source file combines environment access, network access, and code or shell execution; review context before blocking.
dist/ops/login.jsView on unpkg · L1071import { createHash, randomBytes } from "node:crypto";
L2: import { spawn } from "node:child_process";
L3: import { createServer } from "node:http";
L4: import { writeSession } from "./auth.js";
...
L6: export async function browserLoopbackLogin(options = {}) {
L7: const state = randomBase64url(24);
L8: const codeVerifier = randomBase64url(48);
...
L39: headers: { "content-type": "application/json" },
L40: body: JSON.stringify({ code, code_verifier: codeVerifier }),
L41: });
...
L101: function openInBrowser(url) {
L102: const command = process.platform === "darwin"
Findings
4 High3 Medium5 Low
HighChild Processdist/ops/auth.js
HighSame File Env Network Executiondist/ops/login.js
HighSandbox Evasion Gated Capabilitydist/ops/login.js
HighRuntime Package Installdist/ops/workspace.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowNo License
Affected versions and remediation
This report applies to @withfudge/design@2.0.7.
See all versions of @withfudge/design for other recorded verdicts.
- Review the evidence and your use of @withfudge/design@2.0.7 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.