haipipe-individual-inference-report

name	haipipe-individual-inference-report
description	Per-individual prediction-interpretation report: loads one individual's data + recent CGM, hits the deployed prediction endpoint, then asks Claude (via claude_agent_sdk) to compose a dual-layer Report -- structured JSON + natural-language text -- for an audience persona (patient / clinician / etc.). Trigger: individual report, prediction interpretation, generate patient message, /haipipe-individual-inference-report.
argument-hint	--individual <id> --persona <name_or_path> [--endpoint-url URL] [--model X]
allowed-tools	Bash, Read
metadata	{"version":"0.1.0","last_updated":"2026-05-31","summary":"Per-individual prediction-interpretation report."}

Skill: haipipe-individual-inference-report

Per-individual prediction → interpretation → audience-tailored report.

  📥 individual data        🌐 endpoint prediction       🤖 LLM compose
  (parquet)              (haipipe-end-deploy-local)   (claude_agent_sdk)
       │                          │                          │
       └─ ctx ───┬───── forecast ─┴───── system_prompt ───────┘
                │                       (persona)
                ▼
           📨 Report{json, nl}
              + telemetry

Sibling progression in task/4_individual/: