skills/peterbamuhigire/business-plan-skills/ecommerce-unit-economics-and-cross-border-margin-model/Gen Agent Trust Hub
ecommerce-unit-economics-and-cross-border-margin-model
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process sensitive external business data, which creates a standard surface for indirect prompt injection.
- Ingestion points: Data is pulled from external company records, logistics systems, and payment providers as specified in the 'Required Inputs' section of
SKILL.md. - Boundary markers: The skill mandates the creation of an 'assumptions register' where every input must be tagged by source type (e.g., official source, company data) and confidence level.
- Capability inventory: The skill requires read and calculation access. It explicitly prohibits changing live prices, discounts, tax settings, or payments without separate authorized review.
- Sanitization: The workflow requires reconciling orders to refunds and settlements and verifying every load-bearing input against named sources, which acts as a manual validation layer.
- [SAFE]: The skill incorporates explicit safety constraints in the 'Capability Contract' section, specifically warning the agent against making production mutations or reaching financial conclusions without professional review. This limits the risk of the agent taking autonomous actions on live e-commerce infrastructure.
Audit Metadata