konnect-monetization

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill architecture contains an attack surface for indirect prompt injection through external data ingestion.
  • Ingestion points: In SKILL.md, step 2 of the workflow instructions directs the agent to "extract the tier table from the raw HTML yourself" when analyzing published pricing pages.
  • Boundary markers: The instructions do not provide the agent with specific delimiters or boundary markers to isolate untrusted data extracted from HTML content from the agent's core operational logic.
  • Capability inventory: The skill leverages the kong-konnect MCP server and official REST API endpoints (e.g., global.api.konghq.com) to perform write operations in a production billing environment, such as creating meters, plans, and subscriptions.
  • Sanitization: There are no explicit requirements or procedures defined for the agent to sanitize, validate, or filter content retrieved from external websites before using it to generate billing models and entities.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 06:36 PM
Security Audit — agent-trust-hub — konnect-monetization