registry / AI-agent capability abuse / pennyrouter / 0.3.34

pennyrouter@0.3.34

Install and manage PennyRouter local coding-agent integrations.

AI Security Review

scanned 11d ago · by lpm-firewall-ai

The package has an explicit CLI setup flow that mutates Claude Code configuration, which warrants a warning under the agent capability policy. The inspected behavior does not establish malicious execution or an install-time attack.

Static reason

One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source

Trigger

The user runs the pennyrouter install command.

Impact

Claude Code runs the configured status line. No confirmed credential theft, broad control hijack, or npm install-time mutation was found.

Mechanism

The installer configures a statusLine command and writes Claude-specific launch settings; the status line fetches status data from the configured gateway.

AI rationale

Evidence

src/cli.jssrc/launch.jssrc/statusline.js

Network endpoints1

api.pennyrouter.com

Decision evidence

public snapshot

AI called this Suspicious at 87.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

The manifest defines a CLI entry point and only start/check scripts; it has no install lifecycle hook.
Running the explicit install command configures Claude Code's status line and persists the resulting settings.
The status line runs a package-managed script and preserves an existing user status line for restoration.

Evidence against

The flagged agent configuration changes are tied to the user's explicit install command, not npm installation.
CodeGPT setup is manual and reports no JetBrains settings mutation.
The inspected status line code requests data from the configured PennyRouter gateway; no credential theft or unrelated recipient was established.

Behavioral surface

Source

ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell

Supply chain

HighEntropyStringsUrlStrings

ManifestNo manifest risk signals triggered.

Source & flagged code

7 flagged

src/launch.test.jsView file

1import assert from "node:assert/strict"; L2: import { spawnSync } from "node:child_process"; L3: import fs from "node:fs"; ... L53: async function main() { L54: if (!(["darwin", "linux"].includes(process.platform))) { L55: console.log("launch.test.js: skipped on unsupported platform"); ... L59: const root = fs.mkdtempSync(path.join(os.tmpdir(), "pr-launch-test-")); L60: const previous = Object.fromEntries(ENV_KEYS.map((key) => [key, process.env[key]])); L61: const home = path.join(root, "home"); ... L67: }, null, 2)}\n`; L68: const nativeCodex = 'model = "native-model"\n\nmodel_provider = "pennyrouter"\n\n[model_providers.pennyrouter]\nname = "PennyRouter"\nbase_url = "http://localhost:8400/v1"\n'; L69: fs.writeFileSync(path.join(home, ".claude", "settings.json"), nativeClaude);

Medium

Install Persistence

Source writes installer persistence such as shell profile or service configuration.

src/launch.test.jsView on unpkg · L1

src/harnesses/codegpt.jsView file

1import { spawnSync } from "node:child_process"; L2: ... L52: L53: // The record carries the base URL this install actually wrote, so a private/self-hosted L54: // gateway is named here instead of the hosted default the user never configured. ... L58: console.log(" Open CodeGPT settings and remove or replace the PennyRouter provider:"); L59: console.log(` - Base URL: ${record?.changes?.base_url || "https://api.pennyrouter.com/v1"}`); L60: console.log(" - API Key: pr-..."); ... L66: const commands = L67: process.platform === "darwin" L68: ? [["pbcopy"]] L69: : process.platform === "win32"

High

Sandbox Evasion Gated Capability

Source gates dangerous network, credential, or execution behavior behind CI, host, platform, time, or geo fingerprint checks.

src/harnesses/codegpt.jsView on unpkg · L1

src/statusline.jsView file

•Runtime or CLI code links a write operation to an explicit external AI-agent control path: tatusline (all lines) |] penny segment (penny / penny (sub) // / penny (byok) / penny off). the wrapped statusline's first line gets the penny suffix; // any further lines (e.g. a context bar) pass through verbatim. // never blocks on the network: reads a per-session cache and refreshes it detached. import { readfilesync, mkdirsync, writefilesync, statsync } from "node:fs"; import { spawn, spawnsync } from "node:child_process"; import { join, basename } from "node:path"; import os from "node:os"; const statedir = join( process.env.xdg_state_home || join(os.homedir(), ".local", "state"), "pennyrouter", ); const cachedir = join(statedir, "statusline-cache"); const titlesessionfile = process.env.pe

High

External Ai Agent Control Surface Mutation

Runtime or CLI source writes behavior-bearing configuration into a user or project AI-agent control surface.

src/statusline.jsView on unpkg

src/launch.jsView file

•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation. export function pennydatadir() { if (process.platform === "win32") { return join(process.env.localappdata || process.env.appdata || homedir(), "pennyrouter"); } return join(process.env.xdg_data_home || join(homedir(), ".local", "share"), "pennyrouter"); } export function pennyruntimedir() { return join(pennydatadir(), "runtime"); } export function pennycomparedir() { return join(pennydatadir(), "compare"); } export function pennybindir() { if (process.platform === "win32") return join(pennydatadir(), "bin"); return join(homedir(), ".local", "bin"); } export function pennyshimpath() { return join(pennybindir(), process.platform === "win32" ? "penny.cmd" :

High

Trigger Reachable External Ai Agent Control Surface Mutation

Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.

src/launch.jsView on unpkg

standalone/claude-proxy.shView file

•path = standalone/claude-proxy.sh kind = build_helper sizeBytes = 8224 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

standalone/claude-proxy.shView on unpkg

src/mcp-register.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.

src/mcp-register.js#virtual:normalized:round1View on unpkg

src/cli.jsView file

•matchType = previous_version_dangerous_delta matchedPackage = pennyrouter@0.3.17 matchedIdentity = npm:cGVubnlyb3V0ZXI:0.3.17 similarity = 0.710 summary = stored previous version shares package body but lacks this dangerous source file

High

Previous Version Dangerous Delta

This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.

src/cli.jsView on unpkg

Source & flagged code

7 flagged · loading source

src/launch.test.jsView file

Medium

Install Persistence

Source writes installer persistence such as shell profile or service configuration.

src/launch.test.jsView on unpkg · L1

src/harnesses/codegpt.jsView file

Findings

5 High5 Medium4 Low

HighSandbox Evasion Gated Capabilitysrc/harnesses/codegpt.js

HighExternal Ai Agent Control Surface Mutationsrc/statusline.js

HighTrigger Reachable External Ai Agent Control Surface Mutationsrc/launch.js

HighSemantic Analysis Limitedsrc/mcp-register.js#virtual:normalized:round1

HighPrevious Version Dangerous Deltasrc/cli.js

MediumNetwork

MediumEnvironment Vars

MediumInstall Persistencesrc/launch.test.js

MediumShips Build Helperstandalone/claude-proxy.sh

MediumStructural Risk Force Deep Review

LowScripts Present

LowFilesystem

LowHigh Entropy Strings

Affected versions and remediation

This report applies to pennyrouter@0.3.34.

See all versions of pennyrouter for other recorded verdicts.

Review the evidence and your use of pennyrouter@0.3.34 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: 2026-10-11 19:03:49.156Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.