paystack-disputes

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface
  • Ingestion points: The skill retrieves external data such as dispute lists, dispute details, and transaction histories from the Paystack API (referenced in SKILL.md via endpoints like /dispute and /dispute/:id).
  • Boundary markers: The skill does not implement delimiters or explicit instructions to ignore potentially malicious content embedded within the data retrieved from the API.
  • Capability inventory: The skill has the capability to perform network operations (GET, PUT, POST) to the Paystack API, including updating dispute statuses and resolving chargebacks.
  • Sanitization: No sanitization or validation logic is present for the data ingested from the external API before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 06:29 AM