excalidraw-diagrams

Warn

Audited by Socket on Feb 19, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

[Skill Scanner] Download or install from free hosting/deployment platform detected This skill README is coherent with its stated purpose: instructing an agent how to create Excalidraw diagrams and produce .excalidraw files. I found no evidence of obfuscated code, hidden backdoors, or attempts to harvest credentials. The real risks are ordinary operational: writing files to arbitrary paths, executing local Python code via sys.path insertion in examples, and advising uploads to an external MCP/excalidraw.com endpoint which could expose potentially sensitive diagram content publicly. Recommend: validate any MCP URL before use, avoid importing helper code from untrusted local paths, and warn users (as the README does) before uploading sensitive diagrams. Overall, this is not malicious but carries typical safety/privacy considerations when writing or uploading diagrams. LLM verification: The skill fragment is coherently aligned with its stated purpose of guiding Excalidraw diagram creation and optional file export. It includes external dependency references (Excalidraw MCP) that are legitimate but introduce external trust considerations. No sensitive data handling, credential usage, or suspicious data flows are evident in the fragment itself. Given the static nature of the provided content, the risk is low but includes a dependency on an external MCP service which should be revi

Confidence: 80%Severity: 75%
Audit Metadata
Analyzed At
Feb 19, 2026, 02:30 PM
Package URL
pkg:socket/skills-sh/sammcj%2Fagentic-coding%2Fexcalidraw-diagrams%2F@9657a9ce2790c1f99f0f99a692f6e3f602b419f3