paystack-splits

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill contains documentation and helper-based code snippets for managing multi-party payment splits via the Paystack API. It does not include standalone executable files, shell commands, or obfuscated instructions.
  • [INDIRECT_PROMPT_INJECTION]: The skill handles untrusted data inputs such as split names, email addresses, and subaccount IDs, creating a standard surface for indirect prompt injection.
  • Ingestion points: External data enters through parameters in the paystackRequest function calls documented in SKILL.md (e.g., name, email, subaccount).
  • Boundary markers: None are explicitly used in the documentation's code snippets to differentiate between instructions and data.
  • Capability inventory: The skill facilitates network requests (POST, GET, PUT) to the Paystack API endpoints.
  • Sanitization: The provided code snippets use JSON.stringify for basic data serialization, which does not provide protection against prompt-based instruction injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 06:29 AM