Stockbee Episodic Pivot Analyzer classifies candidates after a catalyst such as earnings, a guidance raise, M&A, an FDA decision, an analyst action, a contract or a product launch. It scores catalyst quality together with gap and range expansion, volume shock, neglect context, liquidity and the risk distance to the EP-day low.
Each candidate gets a state (actionable Day 1, Day 1 watch, delayed EP watch, catalyst watch or reject), an EP type and flags for handoff to pead-screener. It does not find news itself and does not give investment advice.
What You Get
- A stockbee_episodic_pivot JSON report with scores and reasons
- A Markdown candidate report
- State and EP type for each candidate, such as EARNINGS_EP or FDA_EP
- Risk to the EP-day low and PEAD handoff and delayed-EP flags
Common Use Cases
- Sorting a list of earnings and guidance events into Day 1 and delayed watchlists
- Combining catalyst context with momentum burst price and volume output
- Passing strong earnings candidates on to a post-earnings drift screener
What Makes It Unique
It scores the news and the price reaction together, and separates names to act on at once from delayed EP names not to chase on Day 1. The catalyst must be supplied, because the skill does not search for news.
Under The Hood
A Python script reads JSON inputs you provide and writes JSON and Markdown reports to a reports folder. Optional price enrichment calls the FMP API over the network with a key from the environment, through a bundled client loaded only for live runs, with a cap on API calls. The SKILL.md has the model run the script from the shell.