Portfolio weights with a traceable solver history.
What it does
Computes portfolio weights from covariance and expected-return data using Conjugate Gradient.
When to use it
Use for mean-variance portfolio optimization with Conjugate Gradient instead of the legacy Neumann series.
How to use it
Give it portfolio data directly or through stored inputs; it computes weights and stores them with solver provenance in trading-risk.
Uses
AG
RU
Latest review
What you need · 2
Uses neural-trader, checking for it and installing it if missing.
Requires the ruflo MCP memory tool to store the resulting portfolio weights and provenance.
Access · 2
This skill
AgentDB
Write
Stores portfolio weights in AgentDB
neural-trader
Execute
Installs neural-trader if missing