israeli-consumer-claim-kit
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user narratives regarding consumer disputes to populate demand letter templates. Ingestion points: User-provided dates, transaction types, and event descriptions collected in Step 1 of SKILL.md. Boundary markers: The skill includes a clear legal disclaimer in both English (SKILL.md) and Hebrew (SKILL_HE.md) stating that outputs are automatically generated and not legal advice. Capability inventory: The skill drafts text and performs mathematical calculations via a local script; it has no network access, file-writing capabilities, or persistent storage operations. Sanitization: No explicit input sanitization is performed, relying on the model's inherent guardrails and the structured nature of the demand letter tracks.
- [SAFE]: The skill references several external resources for legal information including he.wikisource.org, gov.il, consumers.org.il, and kolzchut.org.il. These are established Israeli government, educational, or non-profit legal resource domains and are considered safe for informational purposes.
Audit Metadata