finances-ingest
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
scripts/ingest.py,scripts/sync_moneta.py, andscripts/backfill.py) using theuvpackage manager. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface because it ingests untrusted financial data that could contain embedded instructions.
- Ingestion points: Processes external CSV files from
~/Documents/finances/source/(BofA, Strike, River, Robinhood, and Copilot exports). - Boundary markers: There are no documented boundary markers or instructions to the model to ignore potential instructions embedded within the financial transaction descriptions or metadata.
- Capability inventory: The skill uses
uv run pythonto execute scripts that have file system access (read/write) within the ledger directory. - Sanitization: No sanitization, escaping, or validation of the CSV contents is mentioned in the skill description or command usage.
Audit Metadata