The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's installation instructions direct the user or agent to clone an external repository (https://github.com/drasimwagan/mdv) and execute code via npm install and npm run build. This involves running unverified code from a third-party source not listed as a trusted vendor.
[EXTERNAL_DOWNLOADS]: The skill facilitates downloading code from a non-authoritative repository (drasimwagan/mdv) which is not associated with the primary author's verified infrastructure or recognized trusted organizations.
[COMMAND_EXECUTION]: The documentation provides multiple shell commands including git, npm, and node for setup and operation. Executing these commands in a non-restricted environment could allow for arbitrary code execution if the downloaded repository is compromised.
[INDIRECT_PROMPT_INJECTION]: The MDV tool is designed to ingest and process external data files (CSV and JSON) to generate visualizations, creating a vulnerability surface for indirect prompt injection.
Ingestion points: The data property in the YAML front-matter and inline data within fenced chart blocks in SKILL.md.
Boundary markers: No delimiters or specific instructions to ignore embedded commands are present in the documentation.
Capability inventory: The skill utilizes a CLI (mdv-cli) to render and export documents, which reads from these ingestion points.
Sanitization: No data validation or sanitization steps are documented for the incoming external files.

mdv-markdown-visualization