sharepoint-list-insight-report-generator

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • Indirect Prompt Injection Surface: The skill ingests data from SharePoint list records to generate business insights and statistics. Ingesting untrusted data into the agent's context is a potential vector for indirect prompt injection, where malicious content in a record could attempt to influence the agent's output. \n
  • Evidence: SKILL.md Step 3, 'Analyze all accessible records within the confirmed scope'. \n
  • Sanitization: Step 5 mandates treating fields as untrusted and escaping JSON characters. \n
  • Capability Inventory: Uses SharePoint connector tools to read list data and write files. \n- Dynamic HTML and JavaScript Generation: The skill generates an interactive HTML report including JavaScript from data retrieved at runtime. Building executable files from user-controlled data is a consideration for Cross-Site Scripting (XSS). \n
  • Evidence: SKILL.md Step 5, 'Produce a single HTML file with all CSS and JavaScript embedded inline'. The skill includes mitigations such as using textContent and document.createElement instead of innerHTML. \n- External Dependency (Chart.js): The skill utilizes a third-party charting library loaded from a CDN. Loading external scripts can pose a supply chain risk if the source is compromised. \n
  • Evidence: README.md and SKILL.md Step 5, 'Chart.js (version 4) may be loaded from an exact versioned CDN URL'. The skill mitigates this by requiring Subresource Integrity (SRI) and crossorigin="anonymous". \n- CSV Formula Injection Mitigation: The skill includes logic to sanitize data exported to CSV format. This prevents potential formula injection attacks when the reports are opened in external spreadsheet software. \n
  • Evidence: SKILL.md Step 5, 'For CSV export, neutralize spreadsheet formulas by prefixing an apostrophe...'. \n- Storage Permission Validation: The skill ensures that report files are stored with appropriate access controls. This check prevents data exposure by ensuring that the generated report's storage location respects the original data's privacy constraints. \n
  • Evidence: SKILL.md Step 6, 'Verify that the configured destination's audience is no broader than the source list...'.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 11:56 PM
Security Audit — agent-trust-hub — sharepoint-list-insight-report-generator