skills/holistics/skills/viz/Gen Agent Trust Hub

viz

Pass

Audited by Gen Agent Trust Hub on Nov 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill serves as a guide for authoring visualization blocks within the Holistics platform. It contains legitimate technical documentation and examples without malicious instructions.
  • [EXTERNAL_DOWNLOADS]: The skill references official Holistics GitHub repositories (holistics/ai, holistics/aml) and documentation (docs.holistics.io) for updates and further reading. These are trusted vendor-controlled sources.
  • [INDIRECT_PROMPT_INJECTION]: The skill documents the MarkdownViz feature, which allows rendering data-driven HTML/Markdown. This functionality represents an attack surface where untrusted data could influence rendering, but it is documented as a core feature of the platform.
  • Ingestion points: Processes data from the fetch_dataset tool and user-supplied AQL expressions.
  • Boundary markers: Employs heredoc delimiters like @md and @aql to separate code/templates from instructional text.
  • Capability inventory: Uses the execute_viz tool to render the generated visualization code.
  • Sanitization: Relies on the platform's internal AML compiler and validation tool (holistics aml validate) to enforce structure and safety.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 20, 2026, 09:59 AM