Security audit

Catalyst Calendar

Security checks across malware telemetry and agentic risk

Overview

The skill is a disclosed financial event calendar, but it can automatically create persistent IMA reminder notes without a clear opt-in step.

Install only if you want a finance-focused catalyst calendar that may save reports in your workspace and create IMA reminder notes. Before using it, consider narrowing triggers to market or investing requests and requiring confirmation before any IMA note is created.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Description-Behavior Mismatch

Medium

Confidence: 89% confidence
Finding: The skill's documented primary output is a local catalyst-calendar report, but it also instructs the agent to create IMA reminder notes for high-impact events. This is a scope expansion with side effects beyond the declared artifact, and it can cause unexpected external state changes or data synchronization without explicit user consent.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger phrases are broad enough to match common conversational requests such as asking what important events are coming up next week or this month. This can cause the skill to activate unexpectedly, pulling in web/tool-based behavior when the user may have intended a general discussion, which increases the risk of incorrect routing, unintended data retrieval, or prompt-surface expansion.

Natural-Language Policy Violations

Medium

Confidence: 78% confidence
Finding: The README presents the skill entirely in Chinese and does not document any language-selection behavior or justified locale limitation, which can cause the skill to respond in a language the user did not request. In a multi-skill environment, this may degrade user understanding, increase misexecution risk, and make consent or confirmation prompts less reliable, though it is not a direct code-execution issue.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The condition "需要时间维度的事件前瞻时触发" is ambiguous and lacks operational boundaries, so the skill may activate for a wide range of unrelated forecasting requests. In this skill, ambiguous triggering is more concerning because activation leads to file creation and possibly IMA note creation, increasing the chance of unintended side effects.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The condition "需要时间维度的事件前瞻时触发" is ambiguous and lacks operational boundaries, so the skill may activate for a wide range of unrelated forecasting requests. In this skill, ambiguous triggering is more concerning because activation leads to file creation and possibly IMA note creation, increasing the chance of unintended side effects.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill writes a markdown report into the working directory without disclosing that a file will be created. Silent file creation is a meaningful side effect because it modifies the user's environment, may overwrite expectations about workspace cleanliness, and can create persistence the user did not request.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill directs the agent to create reminder notes in IMA for high-impact events without informing the user about external note creation or synchronization. This is dangerous because it causes actions outside the immediate chat/workspace boundary, potentially exposing user interests or creating persistent reminders in connected systems without consent.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.