markmap
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches necessary rendering libraries from jsDelivr, a well-known CDN, including d3 and markmap-related components.
- [COMMAND_EXECUTION]: The
scripts/render.shscript executesnpxto download and run themarkmap-cliutility for creating offline HTML mindmaps. - [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes external user content to build mindmap hierarchies. * Ingestion points: User-provided notes, architecture documents, or brainstorming text. * Boundary markers: The skill includes explicit instructions in
SKILL.mdto treat input documents strictly as data and never as instructions to execute commands or reveal secrets. * Capability inventory: The skill can perform file writes for.mdand.htmldeliverables and execute therender.shshell script. * Sanitization: Theassets/template.htmlfile includes a runtime sanitizer that usesDOMParserto strip dangerous tags likescript,iframe, andobjectfrom the content before rendering. - [DYNAMIC_EXECUTION]: The skill generates interactive HTML artifacts that utilize JavaScript to transform Markdown JSON into a visual SVG mindmap at runtime.
Audit Metadata