funding-signal-line
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates on a logic-only basis to transform structured company data into natural language sentences. It does not include executable code, remote downloads, or dangerous system commands. The instructions are focused on data accuracy and output formatting.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external company and funding records.
- Ingestion points: Company and funding records fetched via the workspace's company-data source as described in SKILL.md.
- Boundary markers: The skill uses a deterministic wrapper and length limits to isolate and constrain model-generated content.
- Capability inventory: Limited to writing four specific output columns (funding_line, funding_clause, funding_evidence_url, funding_confidence).
- Sanitization: Implements domain equality guards, equity-stage filtering, and recency filtering in code to validate data prior to model processing.
Audit Metadata