Back to skill

Security audit

Generate Minecraft 3D build plan

Security checks for vulnerabilities and agentic risk

Overview

The skill’s remote API behavior matches its stated purpose, but its install instructions ask users to install mutable third-party packages, including a global npm package and an apparently unused Python package.

Review the install path before using this skill. Prefer an isolated environment, avoid elevated privileges, pin dependency versions where possible, and do not send sensitive prompts, private image URLs, secrets, or personal data to the remote API unless you accept that third-party exposure.

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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:8
Finding

Unpinned and Unnecessary Third-Party Dependencies

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 8-15
Vulnerability Type: Unpinned third-party dependencies and unnecessary global package installation
Risk Level: Medium

Vulnerable Code

yaml
dependencies:
  npm:
    - "@aiagenta2z/onekey-gateway"
  python:
    - "ai-agent-marketplace"
installation:
  npm: npm -g install @aiagenta2z/onekey-gateway
  python: pip install ai-agent-marketplace

Technical Analysis

The Skill instructs users to install two third-party packages without exact version or integrity pinning. The npm package is also installed globally. Package names without fixed versions resolve to mutable releases whose contents can change after the Skill has been reviewed.

Neither bundled REST implementation imports these dependencies. The Python script uses standard-library modules, while the TypeScript script uses Node.js globals and the built-in fetch API. Consequently, installing both packages is unnecessary when using the bundled scripts and exceeds the minimum dependency scope required for their declared functionality.

No evidence establishes that either dependency is currently malicious. The vulnerability is the avoidable exposure to package compromise, malicious updates, dependency takeover, and installation lifecycle scripts.

Attack Path

  1. An attacker compromises a listed package, its publisher account, or its package-registry distribution channel.
  2. The attacker publishes a malicious release or modifies an otherwise trusted release.
  3. A user follows the documented installation commands without a pinned version or integrity constraint.
  4. The package manager resolves and downloads the attacker-controlled release.
  5. Package installation or lifecycle hooks execute with the installing user's privileges.
  6. For the global npm installation, attacker-controlled files may be placed in global package and executable locations, increasing the affected sys ...[truncated 608 chars]
Remediation
View remediation

Remediation Suggestions

  1. Remove both dependency requirements when users invoke the included Python or TypeScript REST scripts, because those scripts do not import the listed packages.
  2. Document the OneKey CLI package separately as an optional dependency rather than a requirement for all usage paths.
  3. Pin every necessary dependency to an exact, reviewed version rather than allowing mutable latest-version resolution.
  4. Use package-manager lockfiles and integrity hashes where supported, and verify packages against trusted registry metadata.
  5. Avoid global npm installation. Prefer a project-local installation with a locked dependency tree and invoke the CLI through the local package runner.
  6. Review package provenance, maintainers, lifecycle scripts, transitive dependencies, and published artifacts before approving updates.
  7. Install dependencies without elevated privileges and in an isolated environment or container where practical.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill declares environment-variable and network-dependent behavior but does not define any explicit tool scope such as permissions or allowed-tools. This weakens least-privilege controls and makes it easier for a host agent to grant broader execution or network access than is actually necessary, increasing the blast radius if the skill is misused or modified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs users to send prompts and optional images to an external API using a privileged access key, but it does not disclose the privacy, retention, or third-party handling implications of that transmission. Users may unknowingly upload sensitive prompts, files, or credentials to an external service, and unsafe handling of the API key could also lead to unauthorized use or billing exposure.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script transmits the user-supplied prompt and image URLs to an external service at agent.deepnlp.org, but it provides no explicit user-facing notice at runtime that their inputs will leave the local environment. This creates a data privacy and consent risk, especially if prompts or referenced URLs contain sensitive, internal, or personal information, and the skill context increases concern because remote transmission is central to the tool but under-disclosed in the description.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The manifest says the skill generates a Minecraft 3D scene build plan by calling the Craftsman Agent API, but it does not mention accessing local environment state for credentials. While this is a common implementation pattern for API clients, it is still a local credential-access capability beyond the narrow user-facing purpose described.

Content

No source excerpt is available for this finding.