ClawHub
Back to skill

Security audit

code-to-diagram

Security checks across malware telemetry and agentic risk

Overview

This is a coherent diagram-generation skill, but it can automatically run an unpinned npx renderer and delete Mermaid input files, so it should be reviewed before installation.

Before installing, confirm you are comfortable with this skill running local Node rendering commands. Prefer installing Mermaid CLI yourself from a trusted, pinned source instead of relying on the npx fallback. Use scratch/generated .mmd files or keep backups, because .mmd inputs passed to the renderer may be deleted automatically.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
When `mmdc` is not found, the script falls back to `npx mmdc`, which can trigger package resolution and execution from external sources at runtime. In a security-sensitive or CI environment, this creates a supply-chain risk: an unexpected package version, registry compromise, or path/package confusion could result in arbitrary code execution under the user's privileges.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The help text suggests only intermediate `.mmd` files are deleted, but the implementation can delete a user-supplied `.mmd` passed via `--file`. This can cause silent data loss and, in automated workflows, destroy original user inputs or project files unexpectedly, which is dangerous for integrity even if not directly enabling code execution.

Missing User Warnings

Medium
Confidence
99% confidence
Finding
The script deletes the user-provided `.mmd` file after rendering when `--file` is used and the path ends with `.mmd`, without opt-in or confirmation. In the skill context, users are likely providing source diagrams they expect to keep; silent deletion can lead to irreversible data loss and makes the skill materially more dangerous in automation or batch processing.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical
Code
suspicious.dangerous_exec
Location
scripts/code_to_diagram.js:309