customer-billing-ops
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or security violations were detected within the skill instructions. The skill focuses on legitimate customer support workflows.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill incorporates explicit safety guidelines, instructing the agent never to expose secret keys, full credit card details, or unnecessary personally identifiable information (PII) in its responses.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from support emails and customer complaints, creating a potential surface for indirect prompt injection. This risk is addressed by a logical workflow that requires identity verification and issue classification before any high-impact billing actions are taken.
- Ingestion points: Support complaints, customer emails, and issue tracker content mentioned in SKILL.md.
- Boundary markers: None specified for delimiting untrusted support text.
- Capability inventory: Ability to perform refunds, cancellations, and subscription modifications via integrated billing tools like Stripe.
- Sanitization: Relies on the agent following the provided classification logic to validate requests before execution.
Audit Metadata