Scan a stock market or watchlist to find high-potential investment candidates across the whole market, not limited to any one sector. Use when the user asks to scan the market, shortlist stocks, rank opportunities, build a watchlist, or find candidates that fit value investing, cycle analysis, and reversible turnaround logic. Do not use for deep analysis of one specific stock.
This skill is for broad market scanning.
Its job is to search a market, sector set, or watchlist and produce a ranked shortlist of candidate stocks that deserve further deep research.
This skill must not behave like a single-stock deep-dive skill. Its goal is not to fully value one company. Its goal is to filter the market, reject weak ideas quickly, and surface only the best candidates for deeper work.
The investment framework combines:
This skill covers the whole market. It must not be restricted to resource stocks, state-owned companies, or low-PB names. Those can be valid candidates, but they are only subsets of a larger framework.
Use this skill when the user asks things like:
Do not use this skill when:
In those cases, use the dedicated single-stock deep-dive workflow instead.
The objective is not to always return candidates. The objective is to return only candidates with real research value.
If the market is unattractive, overvalued, crowded, or lacking sufficient evidence, the correct output is:
Never force a shortlist just to satisfy the prompt.
Every candidate must first be classified into one primary type:
If a stock cannot be clearly typed, confidence must be reduced.
Reject or heavily penalize candidates if any of the following is true:
These vetoes matter more than superficial cheapness.
Follow this sequence:
Use this structure unless the user explicitly requests another format:
scripts/engines/radar_scan_engine.py as the scanner entry point.../shared/.no-action as a valid outcome.