Wind Alice Thematic Stock Screening

Thematic stock screening problem

When an engineer asks about AI compute, low-altitude economy, or humanoid robots, the hard part is not finding information. The risk is narrative diffusion: separating core beneficiaries from sentiment spillover, checking whether policy, orders, and earnings form a verifiable path, and judging whether implied valuation is reasonable. This skill turns that question into a structured agent workflow through Wind Alice Agent.

How it works

The host agent composes the user's focus into a natural-language prompt and passes it to the CLI. The service-side agent returns a structured result, usually including:
- Thesis decomposition: upstream, midstream, and downstream drivers, plus policy, order, and earnings signals.
- Data validation: how far the narrative has moved toward fundamentals, with falsification points.
- Beneficiary screening: 5-8 core names, prioritizing logic fit over list length.
- Valuation and trading view: historical percentile, peer comparison, implied expectations, and risk notes.

The host agent should deliver only the agentResult.value content. The full report attachment, if silently downloaded, is for local review and should not be pasted as the user-facing answer.

Boundaries

This skill fits sector investment, industry-chain opportunity research, and concept-stock filtering. Missing market scope, such as A-shares, HK stocks, or US stocks, can reduce precision. It is not investment advice and does not replace compliance or risk controls. If data is limited or the task remains non-terminal, surface the data boundary instead of looping retries or inferring results from logs.