finance-analyze

Pass

Audited by Gen Agent Trust Hub on Jul 5, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its multi-phase orchestration workflow.
  • Ingestion points: The skill collects 30 distinct points of user data (Phase 1), including demographics, income, asset details, and liabilities, which are provided as free-text input by the user.
  • Boundary markers: There are no boundary markers (e.g., XML tags, triple-backticks, or delimiters) or explicit 'ignore embedded instructions' warnings used when these user inputs are passed into the sub-agent prompt templates in Phase 2.
  • Capability inventory: The agent has the capability to write a markdown file (FINANCE-ANALYSIS.md) to the local file system and potentially call other tools via the 'Task' tool mentioned in the instructions.
  • Sanitization: No sanitization, escaping, or validation is performed on the user-provided financial data before it is interpolated into the prompts for the cashflow, debt, investment, retirement, and protection sub-agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 5, 2026, 04:45 PM
Security Audit — agent-trust-hub — finance-analyze