Back to skill

Security audit

Website Analytics

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for website analytics, but it asks users to run unpinned npm packages that could change after review.

Review the Fleets npm packages and publisher before installing, prefer pinning exact versions for both `fleets` and `fleets-mcp`, and run them with only the analytics credentials and environment variables they need.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T08 · Insecure Dependencies

Error
Location
SKILL.md:13
Finding

Unpinned fleets Package Is Automatically Downloaded and Executed

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 13
Vulnerability Type: Unpinned third-party package execution
Risk Level: High

Vulnerable Code

bash
npx -y fleets login --browser   # device flow: opens URL + 8-char code

Technical Analysis

The setup instructions invoke npx with the -y option and do not specify a reviewed package version or integrity constraint. Consequently, npm resolves, downloads, and executes whichever fleets release the configured registry currently serves. The -y option suppresses the interactive installation confirmation.

This creates a supply-chain trust boundary that is not represented by a lockfile, version pin, checksum, or vendored implementation in the audited project. If the package, publisher account, registry response, or resolution environment is compromised, package lifecycle scripts or runtime code can execute with the privileges of the user running the command.

The audit found no evidence that the current fleets package is malicious. The confirmed weakness is the automatic execution of mutable, unpinned third-party code.

Attack Path

  1. An attacker compromises the npm package, its publisher account, or an applicable package-resolution path.
  2. The attacker publishes or serves a malicious release under the fleets package name.
  3. A user follows the documented setup command.
  4. npx -y resolves the current package release without requesting confirmation.
  5. Malicious lifecycle or runtime code executes in the user's environment.
  6. The code may read accessible files and environment variables, including FLEETS_TOKEN, or initiate network connections using the user's privileges.

Impact Assessment

Successful exploitation could provide arbitrary code execution under the account that launches the command. The resulting scope includes that account's readable files, environment variables, network access, and writable resources. Analy ...[truncated 227 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin the CLI to a reviewed, immutable version, for example:
    bash
    npx --yes fleets@X.Y.Z login --browser
    
  • Regularly review and deliberately update the pinned version rather than implicitly accepting the latest release.
  • Where supported, use a lockfile and npm integrity metadata to verify the downloaded artifact.
  • Document the expected npm publisher, source repository, and package identity so users can verify provenance.
  • Prefer installing dependencies in a constrained environment with minimal filesystem access and no unrelated secrets.
  • Avoid exposing FLEETS_TOKEN or other sensitive environment variables while executing newly downloaded package code.
  • Consider removing -y when automatic, unattended package installation is unnecessary.

T08 · Insecure Dependencies

Error
Location
SKILL.md:49
Finding

Unpinned fleets-mcp Package Is Executed as an MCP Server

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 49
Vulnerability Type: Unpinned third-party MCP server execution
Risk Level: High

Vulnerable Code

json
{ "mcpServers": { "fleets": { "command": "npx", "args": ["-y", "fleets-mcp"] } } }

Technical Analysis

The MCP configuration starts fleets-mcp through npx -y without pinning a version or verifying package integrity. Each unresolved installation can therefore retrieve executable content that has changed since the Skill was reviewed. Although the document describes the exposed MCP tools as read-only, that claim constrains the advertised application behavior rather than the operating-system capabilities of the downloaded server process.

A compromised MCP package can execute arbitrary initialization, lifecycle, or server code before any MCP tool is called. It does not need to respect the documented read-only tool interface. No local source is included in the project to verify the server's effective behavior.

The audit found no evidence that the current fleets-mcp package is malicious. The security issue is the unpinned and automatically accepted execution path.

Attack Path

  1. An attacker compromises the fleets-mcp package, its publisher account, or package-resolution infrastructure.
  2. A malicious package release is made available through the registry.
  3. An Agent host loads the documented MCP configuration.
  4. The host runs npx -y fleets-mcp, resolving and downloading the mutable release without confirmation.
  5. The package executes as a local stdio MCP process.
  6. Malicious code accesses resources available to the host account or returns attacker-controlled MCP responses intended to influence subsequent Agent activity.

Impact Assessment

Exploitation could lead to arbitrary code execution with the privileges of the Agent host or user launching the MCP process. Accessible analytics tokens, environment variables, loca ...[truncated 249 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin the MCP server to a reviewed version:
    json
    { "mcpServers": { "fleets": { "command": "npx", "args": ["--yes", "fleets-mcp@X.Y.Z"] } } }
    
  • Use lockfiles and integrity verification where the MCP host and package-management workflow support them.
  • Verify the package's publisher and source repository before approving updates.
  • Run the MCP server in a sandbox with a restricted environment, minimal filesystem permissions, and limited network access.
  • Provide only the credentials required for analytics access; do not inherit unrelated secrets into the MCP process.
  • Review version updates before deployment and maintain an allowlist of approved package versions or artifact hashes.
  • Treat the MCP process as executable third-party code rather than relying solely on its advertised read-only tool semantics.
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The skill instructs users to execute an unpinned package directly from the npm registry via npx -y fleets, which will fetch and run whatever version is current at execution time. If the package is compromised, a malicious version is published, or a dependency is hijacked, this can lead to arbitrary code execution on the user's machine during login/setup.

Content

No source excerpt is available for this finding.