Exposure Coach reads the JSON outputs of other market-analysis skills (market breadth, uptrend, macro regime, market top, follow-through day, themes, sectors and institutional flow) and scores them into one posture summary.
The summary has an equity exposure ceiling from 0 to 100 percent, a growth or value bias, a participation reading, a label of NEW_ENTRY_ALLOWED, REDUCE_ONLY or CASH_PRIORITY, and a confidence level based on how many inputs were present. It is an analytical summary of those reports and not investment advice.
What You Get
- A JSON report with the exposure ceiling, bias, participation, recommendation, confidence and component scores
- A one-page Markdown market posture summary with a rationale
- Lists of provided, missing and unparseable inputs, with warnings
- Timestamped files named exposure_posture_YYYY-MM-DD_HHMMSS in the reports folder
Common Use Cases
- Reading one posture summary at the start of a trading week
- Reconciling conflicting breadth, regime and flow reports
- Re-assessing the posture after a major macro or market event
What Makes It Unique
It is the integration layer for the market-regime skills, and it fails safe: incomplete inputs lower the ceiling and confidence, and invalid upstream reports are excluded without manual repair. Confidence describes input completeness only, and the skill states that it says nothing about freshness, verified quality or prediction accuracy.
Under The Hood
The agent runs a local Python script that reads JSON reports produced by other skills and writes JSON and Markdown posture reports into the reports folder. The prerequisites list an FMP_API_KEY environment variable for the institutional-flow data, so the upstream flow skill reaches a web API. The agent runs shell commands and writes files. Reference files with the scoring rules and the regime-to-ceiling map load when needed.