service-businesses
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from untrusted external sources, creating a potential vector for indirect prompt injection. \n
- Ingestion points: Reads content from public Facebook groups (DEMAND-ALERTS.md), client bills and contracts (BILL-AUDITS.md), chargeback disputes (DISPUTE-RECOVERY.md), and call transcripts (HOSPITALITY.md, TRADE-LEAD-SYSTEMS.md). \n
- Boundary markers: Instructions advise the agent to use minimal quotation, paraphrase inputs, and report unknown facts instead of making assumptions, which helps mitigate accidental obedience to embedded commands. \n
- Capability inventory: Leverages decision models (Jev AI Gateway) and external APIs (Twilio, Stripe, Shopify, Whop) to manage workflows, meaning successful injection could influence real-world business actions. \n
- Sanitization: Mandates human-in-the-loop verification for all outbound artifacts and utilizes the 'anti-slop' standard for content filtering. \n- [NO_CODE]: The skill consists entirely of Markdown instructions and documentation. There are no executable scripts or binary files included, significantly reducing the risk of direct malicious code execution.
Audit Metadata