protheus-reference

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to use external web fetching as a fallback strategy for retrieving table schemas and function documentation, creating a vulnerability to malicious instructions embedded in third-party content.
  • Ingestion points: According to reference.md and sx3-common-fields.md, the agent is instructed to fetch documentation from https://sempreju.com.br/tabelas_protheus/ and the TOTVS TDN documentation via the tdn-lookup skill if local reference files do not contain the requested data.
  • Boundary markers: There are no instructions or delimiters provided to ensure the agent ignores or sanitizes instructions that might be contained within the fetched HTML or JSON content.
  • Capability inventory: The skill provides documentation and examples for high-impact capabilities including network operations (HttpGet, HttpPost, FWCallRest), file system access (FOpen, FWrite), and direct database execution (TCSqlExec, DbSeek).
  • Sanitization: No content validation or sanitization requirements are specified for the external data ingested during the fallback lookup process.
  • [EXTERNAL_DOWNLOADS]: The skill relies on fetching technical specifications from external sources including the official TOTVS TDN documentation and the sempreju.com.br community portal to supplement its local knowledge base.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 11:02 AM
Security Audit — agent-trust-hub — protheus-reference