Loading npm security reports…

warp-agent-mcp@0.19.1: AI-agent control-surface risk | LPM Firewall

registry / warp-agent-mcp / 0.19.1

warp-agent-mcp@0.19.1

MCP server for the Warp freight API, covering the full booking lifecycle: mode compare (one call prices every eligible mode in parallel and returns a decision-complete recommendation with the cost-vs-transit math), quote (van/box-truck/FTL/LTL — LTL split

AI Security Review

scanned 2h ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the package silently modifies Claude Desktop's global MCP configuration. Restarting Claude causes it to execute package code, which reads a local Warp API key and can call the Warp API.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review

Trigger

npm install, followed by Claude Desktop starting or restarting.

Impact

A package-installed MCP server gains automatic execution in Claude Desktop and access to the user's Warp credentials and agent context.

Mechanism

Postinstall persistence in a foreign AI-agent configuration.

Policy narrative

During npm installation, the package runs a postinstall hook that creates or changes Claude Desktop's MCP configuration and points it at its launcher. This occurs without an explicit setup command or user approval. When Claude starts, that launcher executes the package server, reads the user's Warp API key from the home directory, and supports outbound authenticated requests. The automatic foreign-agent configuration mutation is a concrete install-hook abuse chain.

Rationale

The package silently persists itself into Claude Desktop during npm installation, creating automatic execution in a foreign AI-agent control surface. Its subsequent credential access and network capability make this a malicious install-hook abuse path under the policy.

Evidence

package.jsonscripts/postinstall.mjsscripts/run.mjsdist/index.jsdist/client.js~/Library/Application Support/Claude/claude_desktop_config.json~/.warp/config.json

Source & flagged code

7 flagged · loading source

package.jsonView file

•scripts.postinstall = node scripts/postinstall.mjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node scripts/postinstall.mjs

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

scripts/run.mjsView file

9Manifest entrypoint (manifest.bin) carries capability families absent from dist/build output: environment+network, execution+network L9: import { homedir } from "node:os"; L10: import { spawn } from "node:child_process"; L11: import { fileURLToPath } from "node:url"; ... L13: L14: const __dirname = dirname(fileURLToPath(import.meta.url)); L15: // Point at dist/ (built from src/ at version-bump time) instead of the pre-baked ... L22: const homes = [ L23: process.env.HOME, L24: homedir(), ... L30: try { L31: const cfg = JSON.parse(readFileSync(join(h, ".warp", "config.json"), "utf8")); L32: if (cfg.api_key) return cfg.api_key;

High

Entrypoint Build Divergence

Manifest entrypoint contains risky behavior absent from dist/build output.

scripts/run.mjsView on unpkg · L9

•matchType = normalized_sha256 matchedPackage = warp-agent-mcp@0.19.2 matchedPath = scripts/run.mjs matchedIdentity = npm:d2FycC1hZ2VudC1tY3A:0.19.2 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

scripts/run.mjsView on unpkg

scripts/postinstall.mjsView file

5L6: import { readFileSync, writeFileSync, mkdirSync, existsSync } from "node:fs"; L7: import { join } from "node:path"; ... L77: // Check if already configured correctly L78: const existing = config?.mcpServers?.warp; L79: if (existing?.command === warpEntry.command && ... L85: // Add/update the warp entry L86: config.mcpServers = config.mcpServers || {}; L87: config.mcpServers.warp = warpEntry; L88: L89: try { L90: mkdirSync(configDir, { recursive: true });

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

scripts/postinstall.mjsView on unpkg · L5

•matchType = normalized_sha256 matchedPackage = warp-agent-mcp@0.19.2 matchedPath = scripts/postinstall.mjs matchedIdentity = npm:d2FycC1hZ2VudC1tY3A:0.19.2 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

scripts/postinstall.mjsView on unpkg

dist/client.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 2

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

dist/client.js#virtual:normalized:round1View on unpkg

Source & flagged code

7 flagged

package.jsonView file

Lines 16-56json

16 "cursor"

17 ],

18 "homepage": "https://github.com/warpfreight/warp-agent-mcp#readme",

19 "repository": {

20 "type": "git",

21 "url": "git+https://github.com/warpfreight/warp-agent-mcp.git"

22 },

23 "type": "module",

24 "bin": {

25 "warp-agent-mcp": "scripts/run.mjs"

26 },

27 "files": [

28 "bin",

29 "dist",

30 "scripts",

31 "server.json"

32 ],

33 "scripts": {

34 "build": "node scripts/build-widget.mjs && tsc && node scripts/add-shebang.mjs",

35 "dev": "tsc --watch",

36 "postinstall": "node scripts/postinstall.mjs",

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg · L36

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg · L36

37 "check:desc": "node scripts/check-description.mjs",

38 "prepublishOnly": "npm run build && node scripts/check-description.mjs",

39 "test": "node test/refund.mjs && node test/e2e.mjs"

40 },

41 "dependencies": {

42 "@modelcontextprotocol/sdk": "^1.12.1",

43 "zod": "^3.24.4"

44 },

45 "devDependencies": {

46 "@modelcontextprotocol/ext-apps": "^1.7.2",

47 "@types/node": "^25.6.0",

48 "esbuild": "^0.28.0",

49 "typescript": "^5.8.3"

50 },

51 "engines": {

52 "node": ">=20"

53 },

54 "license": "MIT",

55 "bugs": {

56 "url": "https://github.com/warpfreight/warp-agent-mcp/issues"

scripts/run.mjsView file

Lines 1-29javascript

1#!/usr/bin/env node

2/**

3 * MCP launcher — reads API key from ~/.warp/config.json and starts the server.

4 * This runs as the MCP command so it inherits the real HOME and filesystem access,

5 * unlike the bundled server which may run in a sandboxed context.

6 */

7import { readFileSync } from "node:fs";

8import { join } from

scripts/postinstall.mjsView file

Lines 1-25javascript

1/**

2 * postinstall — auto-configures Claude Desktop on install.

3 * Adds the warp MCP entry to claude_desktop_config.json if not already there.

4 */

5

6import { readFileSync, writeFileSync, mkdirSync, existsSync } from "node:fs";

7import { join } from "node:path";

Critical

dist/client.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 2

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

dist/client.js#virtual:normalized:round1View on unpkg