article-taster

• Outcome: Pass
• Latest audit: May 28, 2026 · 1:59 PM UTC
• Version: 1.0.0

Overview

This is a local article-analysis skill with disclosed text/file input and no evidence of credential access, persistence, destructive behavior, or data exfiltration.

Install in a virtual environment, pin and audit dependencies if using it seriously, and only pass article files you intend to analyze. Be aware that the spoiler controls may be unreliable and the AI-detection wording includes harsh/insulting labels that may not be appropriate for user-facing reports.

SkillSpector

Vulnerability Patterns

• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 84% confidence
• Finding: The skill advertises executable Python components plus optional OpenAI/Anthropic access and file-based inputs, but the metadata shown does not declare any permissions or boundaries for file reads or network use. That creates a real security gap: reviewers and hosts cannot tell whether the skill may read arbitrary local files or send article contents to external APIs, which is especially risky because user-supplied articles may contain sensitive or proprietary text.

Intent-Code Divergence

Medium

• Confidence: 85% confidence
• Finding: The filtering logic compares enum string values lexicographically (e.g. 'high' >= 'low') instead of by an explicit severity ranking, so spoiler filtering can behave incorrectly. In a skill explicitly promising spoiler control, this can cause higher-risk content to be insufficiently filtered or filtered inconsistently, undermining user expectations and potentially disclosing sensitive plot information.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

jieba>=0.42.1
scikit-learn>=1.0.0
numpy>=1.21.0

• Confidence: 97% confidence
• Finding: The dependency is specified with a lower bound only, so installs may resolve to different versions over time and across environments. This weakens build reproducibility and can unintentionally pull in newly introduced vulnerable or incompatible releases through the package supply chain.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

jieba>=0.42.1
scikit-learn>=1.0.0
numpy>=1.21.0

• Confidence: 99% confidence
• Finding: Using scikit-learn with only a minimum version allows any later version to be installed, including versions affected by known advisories or future regressions. Because this package has a history of security issues, leaving it unpinned increases exposure and makes deployments non-reproducible.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

jieba>=0.42.1
scikit-learn>=1.0.0
numpy>=1.21.0

• Confidence: 99% confidence
• Finding: The unpinned numpy requirement permits resolution to a broad range of versions, including versions with disclosed vulnerabilities or breaking behavior. In a Python skill that may process untrusted content, uncontrolled dependency drift increases supply-chain and stability risk.

Known Vulnerable Dependency: scikit-learn — 6 advisory(ies): CVE-2020-13092 (scikit-learn Deserialization of Untrusted Data); CVE-2024-5206 (scikit-learn sensitive data leakage vulnerability); CVE-2020-28975 (scikit-learn Denial of Service) +3 more

Critical

• Category: Supply Chain
• Confidence: 92% confidence
• Finding: The requirement allows installation of scikit-learn versions that may include known vulnerabilities, including deserialization and denial-of-service issues. Even if this skill primarily analyzes text, ML/data-processing libraries are often used on attacker-influenced inputs, so vulnerable versions can increase the risk of crashes, data exposure, or unsafe object loading paths.

Known Vulnerable Dependency: numpy — 10 advisory(ies): CVE-2014-1859 (Numpy arbitrary file write via symlink attack); CVE-2021-41495 (NumPy NULL Pointer Dereference); CVE-2021-33430 (NumPy Buffer Overflow (Disputed)) +7 more

Critical

• Category: Supply Chain
• Confidence: 90% confidence
• Finding: The numpy dependency is flagged with multiple advisories, and the current requirement does not exclude affected versions. While many numpy issues require specific local conditions or malformed input paths, a vulnerable scientific-computing library can still enable denial of service, crashes, or unsafe filesystem interactions in some environments.

Static analysis

No suspicious patterns detected.