ux-sentinel

Pass

Audited by Gen Agent Trust Hub on Apr 4, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill implements proactive defenses against indirect prompt injection by instructing the agent to treat all conversation content as untrusted data and to ignore any embedded instructions or skill invocations.
  • Ingestion points: User conversation messages across the session history.
  • Boundary markers: Present; the skill instructions specify that conversation content is data only and directives within it should be ignored.
  • Capability inventory: Notion database interactions via MCP tools (API-post-search, API-post-page, API-patch-page, API-query-data-source).
  • Sanitization: Performs name normalization (lowercase, character removal) on detected concepts.
  • [DATA_EXFILTRATION]: The skill strictly limits data transfer to the user-configured Notion database. It contains explicit guardrails against harvesting credentials or tokens from the local filesystem.
  • [COMMAND_EXECUTION]: All network and system interactions are routed through established MCP tools. The skill explicitly forbids the execution of shell commands or external utilities for API calls.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 4, 2026, 11:07 AM