The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references several external resources for installation and configuration.
Fetches installation scripts from the vendor's official GitHub repository: https://raw.githubusercontent.com/tigerfintech/openapi-python-sdk/master/install.sh and install.ps1.
Downloads the uv tool installer from https://astral.sh/uv/install.sh, which is a well-known service provider in the Python ecosystem.
References official Tiger Brokers documentation at https://docs.itigerup.com and community support via Telegram.
[COMMAND_EXECUTION]: The documentation provides standard command-line instructions for installing and managing the SDK.
Includes commands for package installation via pip, uvx, and conda.
Documents the use of the tigeropen CLI for managing trading accounts and querying market data.
The installation instructions involve piping remote scripts to the shell (curl | sh), which is a common developer practice for the provided vendor resources.
[INDIRECT_PROMPT_INJECTION]: As a trading-oriented skill, it possesses an attack surface where external data could influence agent behavior.
Ingestion points: Processes user-supplied stock symbols, option identifiers, and market data retrieved from the Tiger Brokers API.
Boundary markers: The SKILL.md file defines a clear 'Safety Warning' section that establishes boundaries for the agent's behavior.
Capability inventory: The skill has high-privilege capabilities including place_order, cancel_order, and account asset management across its trade modules.
Sanitization: The instructions mandate a strict safety workflow for live trading, requiring the agent to default to Paper Trading, use preview_order to display fees, and explicitly wait for human confirmation before executing any real-money transactions.

tigeropen