trading-alert-scheduler
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses sensitive financial information, including portfolio positions, account P&L, and margin details for specific accounts (Roth IRA, THK Enterprises) via the IBKR MCP server tools
get_portfolioandget_account_summary.- [COMMAND_EXECUTION]: Instructions rely on an MCP server instance located at a fixed local file path:~/Desktop/tk_projects/ibkr-mcp-server/.- [INDIRECT_PROMPT_INJECTION]: The skill processes unauthenticated market data and news from various web sources (e.g., Yahoo Finance, TradingView) to generate trading signals. - Ingestion points: Stage 1a and 1c pull price action, news catalysts, and market internals from web search results into the agent's context.
- Boundary markers: No specific delimiters or instructions are used to prevent the agent from following potential commands hidden in web news or market commentary.
- Capability inventory: The agent has permissions to write analytics logs to
~/.claude/skill-analytics/last-outcome-trading-alert-scheduler.json. - Sanitization: External content is used directly for regime detection and technical scans without explicit sanitization steps.- [PERSISTENCE]: The skill implements an automated daily execution schedule using a cron expression (
0 7 * * 1-5) to maintain active monitoring and reporting across sessions.
Audit Metadata