Security audit

竞品分析 · 董事会决策报告生成器

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Chinese-language competitive-analysis report template that uses local DOCX/PPTX generation scripts and does not show hidden persistence, credential access, exfiltration, or destructive behavior.

Install only if you want a Chinese-language, template-driven board competitive-analysis workflow. Review and replace the hardcoded Yuxuan Foods data before using it for another company, and provide internal PDFs only when you are comfortable using that business material in generated reports.

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

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (8)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 95% confidence
Finding: The description claims a broader skill that assembles board-level competitive analysis from public web-sourced intelligence and outputs both a DOCX deep report and a 14-slide PPT. The actual code only creates a PPTX presentation using static, prewritten content for a specific example company. There is no evidence of web access, search, crawling, dynamic intelligence collection, document generation for DOCX, or orchestration of two aligned deliverables. While the PPT's subject matter and structure do resemble the declared domain (board competitive analysis, benchmarking matrix, strategic actions, monitoring), the implemented behavior is materially narrower and missing core declared functionality.
Content: No source excerpt is available for this finding.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 92% confidence
Finding: The description promises a broader analytical skill that collects public competitive intelligence via WebSearch and outputs two aligned deliverables: a DOCX report and a board PPT. The supplied code only creates one Word document with fixed, prewritten content for a specific company, using local optional chart files if present. While the document structure strongly matches parts of the declared architecture—data base, competition insight, 9-box benchmarking, strategic decisions, and monitoring—the implemented behavior is materially narrower. It lacks any search, crawling, data retrieval, dynamic company selection, or slide deck generation. Therefore, the description overstates the skill's actual behavior in several important ways, making this a mismatch.
Content: No source excerpt is available for this finding.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: There is a clear mismatch. The declared purpose is a content-generation and strategic-analysis skill that should create two executive deliverables from public competitive intelligence. The supplied code instead performs QA checks on an already-existing local PowerPoint: searching for legacy wording, ensuring certain replacement terms exist, checking slide object bounds, and reporting total slide count. Its primary purpose is presentation validation/editing support, not competitive intelligence research or board-report generation. The resource usage is also different: local PPTX file access via python-pptx rather than WebSearch or multi-document report generation.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The trigger condition is broad enough to activate on many generic requests for market or competitive briefings, which can cause the wrong skill to run and collect, transform, or present sensitive business context without clear user intent. In an agentic environment, over-broad routing increases the chance of unintended file handling, unnecessary web lookup, or misuse of internal documents under the guise of a board-report workflow.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 82% confidence
Finding: The description says the skill is especially for Chinese enterprises and the entire user-facing guidance is written in Chinese, but it does not explicitly offer language or locale choice. That can amount to a locale-policy issue unless the regional limitation is clearly documented as required for compliance or other justified reasons.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The document title and content are entirely in Chinese, and there is no indication that language selection is optional or that the skill is intentionally limited to a Chinese-language or China-specific audience. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy violation.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: This Python skill embeds Chinese-only natural-language content throughout, including the module description, slide titles, labels, and printed output, and it also fixes the font to a Chinese-specific typeface. There is no indication that the user can opt into this locale or choose another language, so the skill effectively forces a specific language by default.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The file header and template instructions indicate the script is specifically structured for Chinese-language report generation, and all embedded output text throughout the script is fixed in Chinese. Because the skill does not provide any opt-in, alternative locale, or justification that it is strictly region-specific tooling, this is a natural-language locale policy concern.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.