Security audit

时事日报分析

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-assisted news-report workflow with local PDF/report utilities and no evidence of hidden persistence, credential access, exfiltration, or destructive behavior.

Before installing, understand that this skill relies on external web searches and will generate reports from public sources, so avoid putting sensitive private context into news queries. Treat generated reports as drafts requiring human review, especially for financial, legal, political, or safety-sensitive decisions. Also note that the bundled PDF script appears to generate a sample-style Chinese report rather than truly converting the provided Markdown input.

A.I.G

Vulnerability Patterns

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
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (13)

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 37)May include surrounding context.
md
| `SKILL.md` | Workflow instructions for AI | AI follows these guidelines |

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The README says the skill should be used 'When a user asks for a news report' and gives the example 'Generate today's international news report.' This does not define clear trigger boundaries, exclusions, or a specific invocation scope, so ordinary requests for news could unintentionally activate the skill.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 90% confidence
Finding: The skill explicitly instructs an AI agent to perform web searches, but the manifest does not declare any tool scope such as allowed-tools or permissions. That mismatch weakens least-privilege controls and can let a host agent invoke broader network/search capabilities than reviewers or policy systems expect.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The markdown explicitly states "Language limitations for non-English/non-Chinese sources," which indicates the skill is constrained to English and Chinese content. This is a natural-language locale/language restriction, and the file does not present it as a user-selectable option or clearly justified region-specific requirement.
Content: No source excerpt is available for this finding.

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 75% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · SKILL.md (reported line 252)May include surrounding context.
md
### Inappropriate Use - ❌ Unattended automated news service - ❌ Real-time trading decisions without verification - ❌ Replacement for professional journalism - ❌ Source of truth for critical decisions without human review

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The template title and all section labels are written in Chinese, which establishes a fixed output language for the skill. There is no visible opt-in, fallback, or indication that the skill is intentionally limited to Chinese-language users or a region-specific use case.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The entire document is written in Chinese and presents its guidance, headings, and usage instructions only in that language, with no indication that users may choose another language or locale. Under the policy for natural-language violations, forcing a specific language without user opt-in is in scope even when it appears in markdown content.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The module docstrings and all user-facing descriptions are written exclusively in Chinese, and the implementation hard-codes Chinese fonts and Chinese report text, indicating the skill is designed to operate in a single language. There is no visible opt-in, language selection, or justification that this is a region-specific tool, which can violate language/locale choice policy.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The command-line interface defines --input as an input Markdown file path, implying the tool will read and convert Markdown content into a PDF. In practice, args.input is never used; the script constructs a fixed sections array with example content and generates a sample PDF regardless of the provided input, which directly contradicts the documented intent of a Markdown-to-PDF tool.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The CLI defaults and generated PDF content use fixed Chinese strings such as the default title, date format, footer text, and subtitle, which force a specific language/locale during normal use. Because no alternative locale or user choice is offered, this is a natural-language policy concern rather than a functional issue.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The verification logic does not match its own stated policy: the docstring says key information requires 2 or more Tier 1-2 confirmations, but the implementation marks Tier 1-2 items as verified by default and accepts Tier 3-4 items with only one matching higher-tier source. In a news-reporting workflow, this weakens source validation and can let inaccurate or manipulated claims be treated as verified, especially if an agent relies on the verified flag when preparing reports.
Content: No source excerpt is available for this finding.

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 78% confidence
Finding: The file states that the AI agent will use web search, collect information from multiple sources, and generate reports, but it does not warn that user requests or contextual data may be sent to external search providers or websites. For a markdown skill description, this is a relevant disclosure because network activity can affect privacy.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 78% confidence
Finding: The module-level description and source definitions present a bilingual/default-language approach, and the data model encodes source language as fixed 'zh' or 'en' values. Under the policy provided, language constraints should be optional or explicitly user-selected unless clearly justified as region-specific.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.