customer-segmentation

Pass

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or security violations were identified during the analysis of the skill's instructions, SQL queries, or TypeScript code snippets.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill facilitates the transfer of customer information to external marketing platforms, which is its primary intended purpose.
  • The syncSegmentToKlaviyo function in SKILL.md targets the official Klaviyo API (a.klaviyo.com) using secure environment variable handling (process.env.KLAVIYO_PRIVATE_KEY) for authentication.
  • The exportSuppressionListForMeta function implements SHA-256 hashing for email addresses using the Node.js crypto module, adhering to privacy standards for third-party audience matching.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides an interface for processing untrusted customer data from databases.
  • Ingestion points include the customers and orders tables mentioned in the SQL and TypeScript snippets.
  • The skill has capabilities to read these records and transmit them to external APIs via network requests.
  • While the provided snippets do not show explicit input sanitization, the logic follows standard data-piping patterns for marketing synchronization rather than risky LLM prompt interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 16, 2026, 11:30 AM