Security audit

doc-convert

Security checks across malware telemetry and agentic risk

Overview

This skill performs disclosed document conversion work, with some install and active-HTML risks users should understand before using it on sensitive or untrusted files.

Install only if you want an agent to run local document conversion tools and create output files for you. Avoid using Marp HTML conversion on untrusted Markdown, review output paths to prevent overwrites, and be aware that npm/npx or system package installs may download external software when dependencies are missing.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The skill instructs users to load Mermaid from a public CDN inside generated slide HTML, which introduces unnecessary network dependency and a third-party code execution path into a local document-conversion workflow. If the generated HTML is opened in a browser, remote JavaScript will execute and could be changed upstream, blocked, or used to leak document context and environment metadata.

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The trigger list includes very broad phrases such as 'document analysis', 'document export', and 'official PDF', which can cause this skill to activate for loosely related requests outside its intended scope. Over-broad activation increases the chance the agent applies this skill's command and tool-selection logic inappropriately, potentially leading to unintended file handling, package-install prompts, or conversion actions on sensitive documents.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill explicitly instructs the agent to execute document conversion immediately when invoked, producing output files without an explicit confirmation step or overwrite safeguard. In an agent setting, this can cause unintended file creation or clobber existing deliverables, especially when output names are predictable (for example, document.docx/document.pdf) and the input path or working directory is user-controlled.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The skill explicitly directs automatic command execution and file creation without requiring user confirmation or warning that the filesystem will be modified. In an agent setting, this raises the risk of unintended writes, overwriting output paths, and executing tooling on attacker-controlled input without adequate trust boundaries.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The documentation promotes the `--html` flag specifically to permit inline HTML tags including `<script>` and then provides a script-loading example, but it does so without any warning about active content execution. In the context of converting potentially untrusted Markdown, this can turn a document conversion flow into a browser-executed script delivery mechanism, increasing XSS-like and data exfiltration risks when the output is viewed.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.