brand-protection-walmart
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill metadata in
SKILL.md(name and description) identifies it as a 'Walmart brand protection toolkit'. However, the core implementation inscripts/detector.pyand the legal templates inscripts/templates.pyuse Amazon-specific identifiers (ASINs) and reference Amazon infrastructure (e.g.,https://brandregistry.amazon.com/). This discrepancy between the skill's identity and its functional implementation is misleading and could result in the incorrect application of brand protection measures. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, specifically product reviews and seller information, which could be used to influence the agent's behavior or poison generated legal documents.
- Ingestion points: Product review content and seller names are ingested and processed in
scripts/detector.pyandscripts/templates.py(ComplaintInfo). - Boundary markers: There are no boundary markers or instructions to isolate external content when generating reports or complaint templates.
- Capability inventory: The skill generates formal legal templates and brand protection reports that are displayed back to the user or agent for further action.
- Sanitization: The skill does not perform any sanitization or validation of the review content or seller strings before interpolating them into the output templates.
Audit Metadata