amazon-one-shot
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources (Amazon product pages) without defining explicit boundary markers or sanitization protocols, creating a risk that instructions embedded in product details or reviews could influence agent behavior.
- Ingestion points: The skill extracts content from Amazon product titles, bullet points, A+ content, and buyer reviews through browser scraping (SKILL.md Phase 1).
- Boundary markers: The instructions do not include delimiters or specific guidance for the agent to ignore or isolate instructions found within the scraped content.
- Capability inventory: The agent has the capability to write multiple markdown files to a local directory (reports/) and perform automated browser navigation based on interpreted data (SKILL.md Phase 2).
- Sanitization: There are no documented steps for sanitizing or validating the scraped text before it is processed by the agent or written to the output reports.
- [DYNAMIC_EXECUTION]: The skill uses browser automation to execute dynamically constructed JavaScript snippets for data extraction from web pages.
- Evidence: The skill relies on browser_evaluate() to execute DOM selection logic such as document.getElementById('productTitle') and document.querySelectorAll() to parse and extract data from the Amazon interface (SKILL.md Phase 1).
Audit Metadata