The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external brokerage APIs and market data feeds, which could be manipulated to influence agent behavior through prompt injection.
Ingestion points: The skill retrieves external data using methods like ib.positions(), ib.accountSummary(), and ib.reqTickersAsync() as documented in reference/connection-patterns.md and reference/trading-patterns.md.
Boundary markers: The instructions do not define clear delimiters or "ignore instructions" warnings when processing the data retrieved from these external sources.
Capability inventory: The skill provides access to high-impact capabilities including ib.placeOrder(), ib.cancelOrder(), and ib.reqGlobalCancel(), allowing the agent to perform financial transactions (in reference/trading-patterns.md).
Sanitization: While a validate_order_for_account function is suggested for checking account restrictions, there is no evidence of sanitization to prevent data from being interpreted as instructions by the LLM.
[EXTERNAL_DOWNLOADS]: The skill encourages the installation and use of several third-party community-maintained libraries and tools that are not from official or trusted vendor repositories.
Evidence: Recommends pip install ib_async (from the ib-api-reloaded/ib_async community repository) and mentions pip install robin-stocks (from jmfernandes/robin_stocks).
Evidence: References various community-created MCP servers such as ArjunDivecha/ibkr-mcp-server, code-rabi/interactive-brokers-mcp, xiao81/IBKR-MCP-Server, and Hellek1/ib-mcp.

ibkr-api-skill