leiloeiro-avaliacao
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data regarding property descriptions and valuation queries, creating a potential surface for indirect prompt injection.
- Ingestion points: The agent accepts property details and market data from users as described in SKILL.md and references/detailed-guide.md.
- Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat external property data as untrusted or to ignore embedded instructions.
- Capability inventory: The skill has the capability to write to a local log file (scripts/governance.py) but lacks network access or system-level modification capabilities.
- Sanitization: No sanitization or validation of user-provided property data is performed before it is processed or logged.
Audit Metadata