| name | help |
| description | Get oriented with the TeamAnalyst plugin and explore available skills. Use when first getting started, checking what analysis capabilities are available, or surveying skills before beginning work. Trigger phrases "start", "get started", "what can you do", "help", "overview", "show skills". |
TeamAnalyst Help
Quick Reference
Main workflow: /reserving-analysis — end-to-end loss reserving (data → methods → selections → projections)
View selection logic: /selection-logic — review LDF and tail factor decision frameworks
Peer review: /peer-review — validate completed analysis against ASOP standards
Step 1: Welcome
Greet the user and briefly introduce the plugin:
TeamAnalyst is an AI-powered actuarial analyst. It walks you through loss reserving — from raw data to development factor selections and ultimate projections — entirely through conversation.
Step 2: Describe the Available Skills
/reserving-analysis — the core workflow. Runs data prep, actuarial methods (Chain Ladder, Initial Expected, Bornhuetter-Ferguson), LDF and tail factor selections, and ultimate projections. Every analysis produces three standard documents: REPORT.md (deliverable), PROGRESS.md (workflow checklist), and REPLICATE.md (reproducibility log)./selection-logic — explains the actuarial LDF and tail factor decision frameworks used during chain-ladder selections./peer-review — reviews a completed analysis against cross-method consistency checks and ASOP standards; produces an advisory findings report without modifying selections.
Step 3: Explain Interaction Modes
When running /reserving-analysis, the user chooses how much control they want:
- Careful — step-by-step review at each stage, full user control over every decision
- Selections Only — automated data prep and calculations, user input only for LDF selections
- Fully Automated — end-to-end execution with no manual intervention
Step 4: Ask How to Help
Ask what the user is working on and guide them to the right skill. Explain that they can invoke any skill by typing its name (e.g. /reserving-analysis).
Reference Sources
If the user asks about specific topics, read the relevant file to answer accurately:
- ASOP standards (13, 23, 25, 36, 41, 43) — see
skills/peer-review/SKILL.md
- Tail curve methods (Bondy, Exponential Decay, McClenahan, Skurnick, etc.) — see
skills/reserving-analysis/agents/selector-tail-factor-ai-rules-based.agent.md
