agentic-fund-orchestration
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides structural guidance for a complex agentic workflow and does not contain malicious code, obfuscation, or unauthorized access patterns.
- [COMMAND_EXECUTION]: The orchestration loop involves placing orders via broker APIs (Alpaca/IBKR) in Step 11. The skill mitigates risks associated with automated execution by mandating human-in-the-loop approval and deterministic risk-management vetoes that reside outside the AI's control.
- [PROMPT_INJECTION]: The skill manages a workflow that ingests external data, presenting an indirect injection surface. * Ingestion points: Market prices from yfinance and macroeconomic data from FRED are ingested in Step 1 (SKILL.md). * Boundary markers: The architecture utilizes a structured JSON state object to separate facts from reasoning and pass data between agents. * Capability inventory: The system contains capabilities for financial data retrieval and trade execution via broker APIs. * Sanitization: Security is enforced through a mandatory human approval step before any execution and a deterministic risk-management layer that can veto agent-proposed trades based on hard caps.
Audit Metadata