ClawHub
Back to skill

Security audit

pdf-processor-for-minimax

Security checks for vulnerabilities and agentic risk

Overview

This PDF-processing skill is purpose-aligned, but users should know it sends PDF content to MiniMax and uses unpinned Python installs.

Before installing, confirm you are allowed to send the target PDFs, including text and embedded images, to MiniMax’s external service. Use an isolated environment, pin dependency versions where possible, and avoid processing confidential documents unless your organization permits that data flow.

A.I.G

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:24
Finding

Unpinned Third-Party Dependencies

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 24-28
Vulnerability Type: Unpinned third-party package installation
Risk Level: Medium

bash
pip install pymupdf  # PDF processing
pip install minimax-client  # MiniMax SDK
pip install pillow  # Image processing

Technical Analysis

The installation instructions retrieve and install the latest available versions of three third-party packages without version constraints, hash verification, or a reviewed lock file. Consequently, the components installed at deployment time may differ from those that were assessed.

If a package, maintainer account, release process, or transitive dependency is compromised, installation or subsequent import can execute attacker-controlled code. The document also does not enforce a trusted package index or verify that the selected MiniMax SDK package is the vendor's official distribution.

Attack Path

  1. An attacker compromises a listed package, one of its transitive dependencies, or its release infrastructure.
  2. The attacker publishes a malicious release under a version accepted by the unrestricted installation command.
  3. A user follows the documented pip install instructions.
  4. The malicious component executes during package installation or when the PDF-processing workflow imports it.
  5. The component operates with the privileges of the Python installation or Skill process and can access data available to that process.

Impact Assessment

Successful exploitation could permit arbitrary code execution with the installing user's or Skill runtime's privileges. Depending on the runtime environment, exposed assets may include processed PDF text and images, the MINIMAX_API_KEY and MINIMAX_GROUP_ID environment variables, filesystem data accessible to the process, and the integrity of generated analysis results.

The finding does not establish that any named package is currently malicious; the risk arises ...[truncated 56 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin every direct dependency to a reviewed, exact version.
  • Resolve and lock transitive dependencies in a committed lock file.
  • Generate and verify package hashes, such as by using pip's --require-hashes mode.
  • Configure an explicit trusted package index or an internally controlled package mirror.
  • Verify the official source, publisher, and package name of the MiniMax SDK before documenting installation.
  • Scan pinned artifacts for known vulnerabilities and repeat the review before dependency upgrades.
  • Install dependencies in an isolated virtual environment or container under a least-privileged account.
  • Example hardened installation pattern:
bash
python -m pip install --require-hashes -r requirements.txt

The corresponding requirements.txt should contain exact versions and approved hashes for direct and transitive dependencies.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill’s workflow explicitly extracts PDF text and images and sends that content to the external MiniMax API for analysis, but the skill description and usage guidance do not clearly warn users that potentially sensitive document contents will leave the local environment. This creates a real privacy and data-handling risk because users may process confidential PDFs without informed consent or without checking whether external transmission is permitted.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

整份技能说明、默认提示词和示例输出均固定为中文，例如默认分析提示为“请分析这个PDF的内容”。文件中没有说明这是特定地区专用技能，也没有提供根据用户偏好切换语言的选项，存在语言/locale 强制约束。

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.