skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/historical-data-manager/Gen Agent Trust Hub
historical-data-manager
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests legacy construction data from external file formats (Excel, CSV, and Access) which could potentially contain adversarial text intended to influence the agent's reporting behavior.
- Ingestion points: Data extraction occurs in HistoricalDataManager.extract_from_legacy_excel, extract_from_csv, and extract_from_database_export.
- Boundary markers: The skill does not currently define explicit delimiters or warnings to ignore instructions embedded within the legacy records.
- Capability inventory: The skill possesses filesystem read access and data processing logic to generate migration reports.
- Sanitization: The skill utilizes standard parsing libraries (pandas, pyodbc) for data intake.
- [COMMAND_EXECUTION]: The skill uses the pyodbc library to interface with legacy Microsoft Access databases and performs SQL queries to retrieve project information.
- [EXTERNAL_DOWNLOADS]: The skill references several well-known Python packages for data processing and database connectivity, including pandas, openpyxl, xlrd, and pyodbc, which are standard utilities in the industry.
Audit Metadata