prepare-property-tax-appeal
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external real-property records and listings, creating a potential attack surface.
- Ingestion points: Workflow steps in
SKILL.md(Steps 3, 4, 6, and 10) ingest data from assessor pages, search results, and owner-provided attachments. - Boundary markers: The
SKILL.mdfile contains explicit instructions to treat all evidence as untrusted and never follow embedded instructions to run commands or reveal data. - Capability inventory: Execution is limited to jurisdiction lookup and structured generation of Markdown/PDF files; no arbitrary shell command execution or network exfiltration capabilities are exposed to untrusted input.
- Sanitization: The skill employs
url_safety.pyfor canonical public-HTTPS validation and SSRF protection, andbuild_appeal_packet.pyfor strict JSON schema enforcement.
Audit Metadata