skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/data-model-designer/Gen Agent Trust Hub
data-model-designer
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided construction data (CSV, Excel, JSON) to generate data models and schemas.
- Ingestion points: Project data and field parameters are accepted via user input or file references (specified in
instructions.md). - Boundary markers: No explicit delimiters are used in the prompt interpolation within
SKILL.md, althoughinstructions.mddirects the agent to validate inputs. - Capability inventory: The provided Python logic is restricted to data transformation; it contains no subprocess calls, file system writes, or network operations.
- Sanitization: The script performs basic type mapping but does not sanitize string inputs used in generated SQL/Mermaid outputs. This poses a negligible risk as the skill only produces text and does not execute the generated code.
Audit Metadata