Lightweight charting skill: powerful styling & direct PNG export

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:37 PM UTC
• Version:
  1.0.13

Overview

This is a coherent chart-rendering skill with an expected local Node runtime, a disclosed external style-tuning page, and scoped saved chart presets.

Install only if you are comfortable with an npm-based local chart renderer. Keep sensitive datasets out of the hosted config page, review generated JSON before saving it, and remember that saved config changes affect future charts of that type.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README explicitly instructs users to visit an external configuration page, generate a config there, and paste the result back into the chat. In an agent workflow, that can cause sensitive business data, chart labels, metrics, or styling metadata derived from private datasets to be disclosed to a third-party site or copied back into a broader conversation context without any warning, review step, or data-minimization guidance.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The instruction tells the agent to save user-provided config JSON into a persistent file under `{baseDir}/config/` without explicitly warning the user that this modifies skill state. That can lead to unintended persistent changes, cross-session tampering of future chart outputs, and abuse if hostile JSON is supplied or if the user did not mean to alter shared defaults.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: These post-handoff instructions explicitly direct the agent to write copied JSON into `skills/data-charts-visualization/config/<chart>_style.json`, a persistent skill config file, again without requiring an explicit user warning or confirmation. In this skill context, chart config is trusted to influence rendering behavior, so persisting unreviewed user-supplied content can poison future renders, overwrite defaults, or create a durable prompt/config injection surface.

VirusTotal

64/64 vendors flagged this skill as clean.