freeStockLIneskill

The name/description match the code: the runtime implements routing, source clients, and endpoints for A‑share data and calls public finance endpoints and free Python libs (AKShare, efinance, Tencent, Sina, etc.). Requiring only python3 is plausible, but the code clearly depends on additional Python packages (requests, efinance, akshare, pandas) that are not declared in the skill metadata.

SKILL.md directs the agent to forward the user's original natural‑language query to the smart-query CLI which will in turn query external public endpoints and libraries. That matches the stated purpose, but it means the user's exact query text is sent to third‑party servers (smartbox.gtimg.cn, suggest3.sinajs.cn, web.ifzq.gtimg.cn, and others) and may cause the agent to fetch announcement PDFs or other remote resources — expected for data collection but a privacy/exfiltration consideration.

There is no install spec despite multiple non‑standard Python dependencies used in the code (requests, efinance, akshare, pandas, efinance/efinance import in history path). The bundle includes .py runtime files (extract/execute on install is not used), but without declared dependency installation the skill may fail or behave unpredictably depending on the host environment. The absence of a requirements list or safe, auditable install steps is an operational / supply‑chain concern.

The skill requests no environment variables, no credentials, and sets session.trust_env=False to avoid inheriting proxy settings. All network calls go to public finance endpoints or libraries invoked locally. No unrelated secrets or system config paths are requested.

always is false and the skill is user‑invocable; openai.yaml allows implicit invocation (normal). The skill does not request persistent system privileges and does not modify other skills or system settings.