Back to skill

Security audit

Docx Analysis

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward DOCX analysis helper, with a supply-chain caution because it asks users to install an unpinned third-party CLI globally.

Install only if you are comfortable using MinerU's CLI for DOCX processing. Prefer a pinned or isolated installation when possible, and avoid processing sensitive documents until you understand where the CLI sends data and how MINERU_TOKEN is stored or used.

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:14
Finding
Unpinned Third-Party CLI Packages Installed Globally## Vulnerability Details **File Location**: `SKILL.md`, lines 14–17 **Vulnerability Type**: Unpinned third-party dependencies and unsafe global installation **Risk Level**: Medium ### Vulnerable Code ```bash npm install -g mineru-open-api # or via Go (macOS/Linux): go install github.com/opendatalab/MinerU-Ecosystem/cli/mineru-open-api@latest ``` ### Technical Analysis Both installation commands resolve mutable third-party package versions at installation time. The npm command omits a version specifier, while the Go command explicitly requests `@latest`. The project provides no lockfile, cryptographic checksum, signature, or immutable source reference with which to verify the installed implementation. The packages are installed globally, exposing the resulting executable through the user's command search path. Package installation may also execute upstream installation logic with the privileges of the invoking user. Consequently, a compromised package registry account, source repository, release artifact, maintainer account, or transitive dependency could introduce code that was not present during this audit. This finding concerns supply-chain integrity. The reviewed file does not itself contain malicious code, and there is no evidence that the named upstream project is currently compromised. ### Attack Path 1. An attacker compromises an upstream maintainer, package registry account, repository, release pipeline, or transitive dependency. 2. The attacker publishes a malicious release that becomes the latest available version. 3. A user or automated agent follows the documented installation instructions. 4. npm or Go resolves and downloads the attacker-controlled release because no reviewed immutable version is specified. 5. Installation-time code or the subsequently invoked `mineru-open-api` executable runs under the installing user's account. 6. The malicious package may access files, environment variables, credentials, and network resources available t ...[truncated 642 chars]
Remediation
## Remediation Suggestions 1. Pin both installation methods to specific, reviewed versions instead of resolving the latest release: ```bash npm install -g mineru-open-api@<reviewed-version> go install github.com/opendatalab/MinerU-Ecosystem/cli/mineru-open-api@vX.Y.Z ``` 2. Publish expected checksums or signature-verification instructions for the approved release artifacts. 3. Prefer project-local or isolated installation over global installation where operationally possible. 4. Review the pinned package, its installation hooks, and its transitive dependencies before approval. 5. Use automated dependency monitoring, but update versions only after security review and integrity verification. 6. Run document processing with least privilege and restrict access to unrelated files, credentials, and network resources. 7. Clarify whether DOCX content is transmitted to external services and document applicable retention and privacy controls before sensitive documents are processed.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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

Static analysis

No suspicious patterns detected.