The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to run Python scripts located in a sibling directory (../findata-toolkit-cn/scripts/views_runner.py). These scripts are used to fetch financial market data such as stock prices and historical data.
Evidence: Found in references/data-queries.md where commands like python ../findata-toolkit-cn/scripts/views_runner.py stock_zh_a_hist are documented.
[EXTERNAL_DOWNLOADS]: The skill guides the agent to install Python dependencies from a local file during environment preparation.
Evidence: references/data-queries.md contains the instruction python -m pip install -r ../findata-toolkit-cn/requirements.txt.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided portfolio data and external financial data, which represents a potential injection surface if the external data sources were compromised.
Ingestion points: User input for holdings (Step 1 in SKILL.md) and external market data via views_runner.py.
Boundary markers: The skill uses structured templates (references/output-template.md) to format results, which provides some structure to the output.
Capability inventory: The skill has access to shell command execution for running diagnostic scripts.
Sanitization: No explicit sanitization or filtering of external data content is described in the prompt instructions.

portfolio-health-check