The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the uv environment manager from a well-known GitHub repository to provision a managed Python runtime for GUI components.
Source: astral-sh/uv releases on GitHub.
Security: The download is verified against pinned SHA256 hashes defined in scripts/weex_gui_bootstrap.py before execution.
[COMMAND_EXECUTION]: The skill executes local system commands for environment management and runtime setup.
Dependency installation: Uses pip to install packages from requirements.lock using the --require-hashes flag for integrity.
Process management: Uses os.execve to relaunch itself under the managed Python virtual environment when required for GUI compatibility.
Local agent: Spawns a background process (weex_vault_agent.py) that acts as an in-memory session store for decrypted vault keys, reducing on-disk exposure.
[PROMPT_INJECTION]: The skill processes untrusted order data from CLI arguments, representing a surface for indirect prompt injection (Category 8).
Ingestion points: Receives order details via --order-json and --tp-sl-json in scripts/weex_trade_guard.py.
Boundary markers: Employs a user_environment_prefix to clearly distinguish between system status information and data-derived summaries.
Capability inventory: Capable of performing authenticated REST API calls and local file/process operations.
Sanitization: Input data is strictly parsed as JSON and validated against expected numeric and categorical values before use.

weex-trader-skill