Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Microsoft MarkItDown
v1.0.0Use MarkItDown to convert various files (PDF, Word, Excel, PPT, images, audio, HTML, CSV, JSON, etc.) to Markdown format for LLM processing and text analysis...
⭐ 0· 10·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name and description (convert many filetypes to Markdown) match the SKILL.md instructions (use the markitdown CLI/Python API). However the skill's name uses 'Microsoft' while the implementation references a third‑party PyPI package 'markitdown' and optional Azure/OpenAI integrations — the Microsoft branding may be misleading and the provenance of the package is not provided.
Instruction Scope
Instructions tell the agent to install and run a third‑party package and optionally to instantiate LLM clients (OpenAI) or call Azure Document Intelligence. Enabling OCR/LLM features implies sending extracted file contents to external APIs; the SKILL.md does not provide any data handling / privacy guidance or declare the credentials/environment variables required for these flows. The agent is also instructed to iterate ZIP contents and process arbitrary files supplied by the user — normal for the task, but increases data‑exfiltration risk when combined with LLM/Azure calls.
Install Mechanism
There is no formal install spec in the registry; the SKILL.md recommends running pip install 'markitdown[all]'. Installing an opaque PyPI package (and its extras) from an unspecified source is a moderate-to-high risk because package provenance, maintainers, and contents are unknown. The SKILL.md also suggests optional extras that will pull additional dependencies (potentially including network‑enabled tools) without documenting them.
Credentials
The instructions reference needing an LLM API key (OpenAI) and Azure Document Intelligence configuration for optional features, yet the registry declares no required environment variables or primary credential. This mismatch means the skill can prompt the agent/user to provide secrets at runtime but gives no upfront justification or scoping for those credentials. Requesting API keys to external providers is plausible for OCR/LLM and Azure features, but the absence of declared env vars and lack of guidance on minimal scopes is a red flag.
Persistence & Privilege
The skill is instruction-only, has no install spec in the registry, does not request persistent 'always' inclusion, and doesn't claim to modify other skills or system-wide settings. It can be invoked autonomously (platform default), which is expected; no elevated persistence is requested.
What to consider before installing
This skill looks like a wrapper around a third‑party PyPI package (markitdown) and optionally forwards content to external LLM/Azure services. Before installing or using it: 1) verify the package source — find the PyPI page or source repository and review the maintainer, recent releases, and code (look for typosquatting and unexpected network/callouts); 2) avoid installing into your system Python — use an isolated virtualenv or sandbox; 3) do not enable OCR/LLM/Azure features or provide API keys unless you trust the provider and understand that file contents will be transmitted off‑host; 4) ask the skill author for a homepage/repository link, a list of dependencies installed by the extras, and checksums for releases; and 5) if you must test it, try it on non‑sensitive sample files first. Additional information that would raise confidence: authoritative homepage/repo, PyPI package page, vendor/owner identity, and an explicit declaration of required environment variables and data‑handling policy.Like a lobster shell, security has layers — review code before you run it.
latestvk973jycb3hqgpn2s5zfjfb7j8584p5rj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.