skills/serenorg/seren-skills/tax/Gen Agent Trust Hub

tax

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill connects to the Kraken API (api.kraken.com) to retrieve trade and ledger history and the SerenDB API (api.serendb.com) to manage database targets and connection strings.- [DATA_EXFILTRATION]: Transmits processed reconciliation summaries and normalized tax records to the vendor's SerenDB platform for storage when advanced features are enabled.- [COMMAND_EXECUTION]: Orchestrates multiple internal Python scripts to perform data processing tasks and uses dynamic module loading in run_pipeline.py to handle normalization logic.- [PROMPT_INJECTION]: The skill ingests untrusted financial data from user-uploaded files, creating a surface for indirect prompt injection. \n
  • Ingestion points: scripts/1099da_normalizer.py and scripts/reconciliation_audit.py (via common.py). \n
  • Boundary markers: The skill processes data as structured CSV or JSON but does not implement explicit natural language boundary markers or instructions to ignore embedded commands. \n
  • Capability inventory: Includes network requests to external APIs, file system write access for output artifacts, and database write capabilities. \n
  • Sanitization: Employs standard JSON serialization and uses parameterized SQL queries via psycopg to prevent injection into database commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 02:44 AM