fabriciq-ontology-cli

Pass

Audited by Gen Agent Trust Hub on Aug 9, 2026

Risk Level: SAFE
Full Analysis
  • Official Platform Interaction: The skill exclusively utilizes official Microsoft Fabric endpoints, such as api.fabric.microsoft.com and kusto.kusto.windows.net, for item management and data exploration.
  • Security Guardrails: The skill implements a mandatory 'Preview & Confirm' workflow (documented in preview-and-confirm.md), which requires agents to present an ASCII proposal or diff and obtain user confirmation before executing write operations.
  • Indirect Prompt Injection Surface: The skill ingests untrusted data from Fabric APIs (ingestion points: getDefinition, schema discovery). It mitigates risks through the aforementioned boundary markers and instructions for sanitization, such as merging schemas to detect property name collisions.
  • Legitimate Data Processing: Base64 encoding/decoding is used for its intended purpose of handling Fabric item definitions. The scripts provided are educational templates rather than hidden execution payloads.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 9, 2026, 08:57 AM
Security Audit — agent-trust-hub — fabriciq-ontology-cli