Drawio Skill

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The skill appears coherent for creating Draw.io diagrams, with expected local CLI/file operations and persistent style presets, and no evidence of hidden exfiltration or credential misuse in the provided artifacts.

This skill looks reasonable to install if you trust the Draw.io desktop app and are comfortable with local file generation. Use local exports for confidential diagrams, review any prompt to run `git pull`, and save or delete style presets only when you intend them to persist across future diagram requests.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

#
ASI02: Tool Misuse and Exploitation
Low
What this means

The agent may run Draw.io commands locally and write generated diagram/export files.

Why it was flagged

The skill intentionally runs a local desktop CLI and writes output files; this is central to the diagramming purpose but users should understand it can create or overwrite files at chosen output paths.

Skill content
Generate `.drawio` XML files and export to PNG/SVG/PDF/JPG locally using the native draw.io desktop app CLI.
Recommendation

Use intentional output paths and review before allowing file overwrites, especially in shared or important directories.

#
ASI04: Agentic Supply Chain Vulnerabilities
Low
What this means

Accepting an update could replace the skill’s instructions or helper files with whatever is in the configured upstream repository.

Why it was flagged

The skill may check a git remote and can self-update only after user consent. This is disclosed, but it depends on the trustworthiness of the configured origin.

Skill content
fetch the latest tag from upstream: `git -C <this-skill-dir> ls-remote --tags origin 'v*'` ... If they say yes, run `git -C <this-skill-dir> pull --ff-only`.
Recommendation

Only approve `git pull` if you trust the repository origin and prefer registry-managed updates when available.

#
ASI06: Memory and Context Poisoning
Low
What this means

Saved presets can affect later diagram generation and may retain provenance such as the source file path.

Why it was flagged

User style presets are persistent local state and can become defaults that influence future diagrams.

Skill content
`~/.drawio-skill/styles/<name>.json` — user presets ... `save` / `looks good` → write candidate to `~/.drawio-skill/styles/<name>.json`.
Recommendation

Save presets only after reviewing them, and inspect or delete files in `~/.drawio-skill/styles/` if a default style behaves unexpectedly.

#
ASI07: Insecure Inter-Agent Communication
Low
What this means

Opening the fallback URL loads an external viewer page that processes the diagram content in the browser.

Why it was flagged

When the CLI is unavailable, the fallback creates an external diagrams.net viewer URL containing the diagram XML in the fragment.

Skill content
Produces a client-side URL — the diagram XML is encoded in the URL fragment ... `https://viewer.diagrams.net/...#R` + urllib.parse.quote(encoded, safe="")
Recommendation

For sensitive diagrams, prefer the local Draw.io CLI export path and avoid sharing or opening encoded viewer URLs unless you trust the viewer environment.