Excalidraw Diagram Generator

• Outcome: Pass
• Latest audit: May 29, 2026, 8:17 PM
• Version:
  1.3.1

Overview

This skill is a disclosed diagram generator that creates local Excalidraw files, with no evidence of hidden execution, credential access, exfiltration, or destructive behavior.

Install this if you want automatic Excalidraw diagram files. Be aware that generic prompts like diagram, flowchart, or mind map may trigger it, and it will create output files in the current working directory, so use it from a folder where those files are expected.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The Excalidraw skill advertises generic triggers such as "diagram" and "flowchart," which are common user requests and can cause the skill to activate in situations where a more appropriate skill or default behavior should be used. In a prompt-driven agent system, overly broad activation increases the chance of unintended prompt injection surface, misrouting, or execution of the wrong skill logic from ordinary language.

Vague Triggers

Medium

• Confidence: 85% confidence
• Finding: The Mermaid skill uses ambiguous triggers like "visualize" and "flowchart," which are broad enough to match many unrelated requests. This can lead to unintended skill invocation, especially in multi-skill environments, creating reliability and security issues by expanding the set of prompts that load specialized instructions unexpectedly.

Vague Triggers

Medium

• Confidence: 83% confidence
• Finding: The Canvas skill includes broad phrases like "mind map" and "visual diagram," which may overlap with normal conversation and other visualization skills. In a shared skill ecosystem, this raises the risk of accidental activation and competing prompt instructions, which can degrade safety boundaries and cause the wrong formatter or workflow to run.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The documented trigger phrases are broad and overlapping across multiple visualization skills, including generic terms like “可视化”, “流程图”, and “思维导图”. In an agent environment that auto-loads skills by trigger text, this can cause unintended activation, misrouting user requests, or unnecessary exposure of skill instructions, which is a real prompt-scope/security weakness even if not overtly malicious.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger list contains broad, common terms such as "diagram", "animate", and generic Chinese terms for drawing/visualization, which can cause the skill to activate in situations the user did not intend. In a skill that also writes files automatically, accidental invocation increases the chance of unexpected local file creation and misrouting of user requests.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The skill explicitly instructs automatic saving to the current working directory and later mandates using a write tool, but it does not require clear prior consent before modifying the local filesystem. This is dangerous because a user may believe they are only requesting content generation, while the skill performs a side effect that creates files in the environment automatically.

VirusTotal

64/64 vendors flagged this skill as clean.