Security audit

Meeting Notes Generator

Security checks for vulnerabilities and agentic risk

Overview

This skill is a meeting-notes helper whose behavior is mostly disclosed and purpose-aligned, but users should treat external AI processing and company-name validation as privacy and accuracy considerations.

Install only if you are comfortable processing meeting transcripts with an external AI service. Redact confidential content where needed, confirm your organization allows third-party AI processing, and manually verify names, numbers, decisions, deadlines, and action items against the original transcript.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Description-Behavior Mismatch

Medium

Confidence: 94% confidence
Finding: The prompt file embeds a company-name verification and business-research workflow that is outside the stated scope of a meeting-minutes summarization skill. Unrelated capabilities increase attack surface, can trigger unintended data handling or hallucinated external-research behavior, and violate least-privilege expectations for users invoking a summarization tool.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: Granting business-research/company-verification behavior to a meeting-minutes skill is unjustified by the skill's declared purpose and can cause the agent to process requests beyond user expectation. In practice this may lead to unauthorized enrichment of meeting content, privacy issues, and misleading outputs based on 'simulated search' rather than grounded source material.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The skill explicitly says it uses an external DeepSeek API to process meeting content, but the description does not warn users that uploaded transcripts may be transmitted to a third party. Meeting records often contain confidential business discussions, personal data, decisions, and action items, so undisclosed external transfer creates a significant privacy and compliance risk.

Static analysis

No suspicious patterns detected.