| name | concentration-risk-assessment |
| description | Use this skill when the user asks about "concentration risk", "which positions need attention", or "position limit breaches". Retrieves thresholds from internal policy documents and applies them to portfolio positions. |
Concentration Risk Assessment
When to Activate
Trigger when user asks about: "concentration risk", "position limits", "which positions need attention", "concentration breach", "portfolio concentration"
Policy-Driven 2-Phase Approach
Phase 1: Retrieve Policy Thresholds
Tool: search_internal_docs
Query: "concentration risk limits", "issuer concentration", "position limits"
Extract from policy:
- Warning threshold: typically 6.5%
- Breach threshold: typically 7.0%
- Any sector-specific limits
Phase 2: Apply to Portfolio Data
Tool: quantitative_analyzer
Query: "Current position weights for SAM [portfolio]"
Calculate: Position weight = Market Value / Total Portfolio Value
Flagging Rules
Apply policy thresholds to flag positions:
- Warning (6.5-7.0%): "CONCENTRATION WARNING — Per Concentration Risk Policy"
- Breach (>7.0%): "BREACH — Immediate remediation required per policy"
- Within limits (<6.5%): No flag needed
Output Format
| Position | Weight | Status | Policy Reference |
|---|
| [Ticker] | X.X% | BREACH / WARNING / OK | Concentration Risk Policy §X |
Total flagged exposure: [sum of warning + breach positions]
Recommendations
- Breach positions: Immediate Investment Committee review required. Recommend reducing to below 6.5% within [timeframe from policy]
- Warning positions: Monitor daily. Place on watch list. Prepare contingency reduction plan
- Multiple flags: Prioritise by breach severity and liquidity
Cite specific policy sections for all threshold references.
Stopping Points
- After Phase 1 (policy thresholds retrieved): confirm thresholds with user before applying
- After Phase 2 (positions flagged): present flagged positions for review before recommending actions
Output
A concentration risk report with flagged positions table, severity classifications, and policy-aligned recommendations.