markstream-install

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to configure a markdown renderer for processing untrusted AI-generated content, creating a potential surface for indirect prompt injection.\n
  • Ingestion points: The agent analyzes repository configuration files (e.g., package.json) and application source code during setup, and the resulting components render dynamic AI-generated text/SSE tokens at runtime.\n
  • Boundary markers: Instructions explicitly require preserving security defaults, such as html-policy="safe" and strict mode for Mermaid diagrams, to prevent execution of malicious embedded content.\n
  • Capability inventory: The skill enables the agent to install NPM packages, modify application entry points, and execute project build/test commands.\n
  • Sanitization: Security relies on the internal sanitization mechanisms of the Markstream library for HTML and diagram rendering.\n\n- [EXTERNAL_DOWNLOADS]: The skill involves the installation of external Node.js packages.\n
  • Details: Recommends installing vendor renderers (e.g., markstream-vue) and well-known peer dependencies such as mermaid, katex, and @antv/infographic from the standard npm registry.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 09:50 AM
Security Audit — agent-trust-hub — markstream-install