Skills
Navigate
skills/serejaris/kimi-skills/interactive-research-report/Gen Agent Trust Hub

interactive-research-report

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting untrusted data from research reports (PDF/Markdown) and data tables (CSV/JSON). The provided utility script templates/utils.js contains functions like showDrill and frame that use .innerHTML to render variables such as title, sub, and value directly into the document. If the input report contains malicious payload, it could lead to Cross-Site Scripting (XSS) in the generated website.
  • Ingestion points: Data enters the agent context via user-provided reports and is then compiled into js/data.js and js/sources.js for the website.
  • Boundary markers: The skill does not implement or recommend explicit sanitization or escaping of the ingested text before rendering.
  • Capability inventory: The skill enables the generation of client-side JavaScript that manipulates the DOM based on input data.
  • Sanitization: Sanitization is absent in the templates/utils.js file, which is the core rendering engine for the generated reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 08:59 AM