Summary
New skill lookthrough-concentration: computes true single-name exposure by looking through funds (ETFs, TDFs, index funds) and aggregating the same company across accounts, share classes, and listings. It produces the number a concentration policy is actually enforced against: lookthrough exposure ÷ total book.
Gap
No current vertical computes name-level exposure through funds. The canonical advisor case: employer stock held directly + ESPP/RSU lots + a broad index fund + a target-date fund — every account-level view says "diversified," while the household owns one company four ways. Rebalance review is blind to this.
What the skill does
- Aggregates by company, not ticker (share classes like GOOG/GOOGL; cross-venue pairs like SAP vs SAP.DE), aliases resolved on both sides so drill-downs match the aggregate
- Country-guards aliases (
ROG in a US fund is Rogers Corp, not Roche); ambiguous tickers stay un-merged
- Normalizes provider weights summing >100%; attributes <100% remainders instead of dropping them
- Uses lookthrough ÷ total book as the cap metric with explicit denominators; flag-vs-gate semantics — an over-benchmark weight points at the name-level breach, never auto-recommends selling a sector or fund
- Per-name drill-down by channel with synthetic shares (position qty × holding weight)
Plus portability guardrails: holdings-file staleness stamps, one-level-only expansion for fund-of-funds, non-equity sleeves excluded from the equity view.
Provenance / fit
Distilled from a production family-office portfolio system I run — method and failure modes only, provider-neutral, markdown-only, styled after thesis-tracker. A complete SKILL.md is drafted and ready. Fits plugins/vertical-plugins/financial-analysis/skills/ (shared core) or the advisors vertical alongside rebalance review.
Next step
Repo forked (sbhavani/financial-services). Happy to open the PR once placement is confirmed — or iterate on scope here first.
Summary
New skill
lookthrough-concentration: computes true single-name exposure by looking through funds (ETFs, TDFs, index funds) and aggregating the same company across accounts, share classes, and listings. It produces the number a concentration policy is actually enforced against: lookthrough exposure ÷ total book.Gap
No current vertical computes name-level exposure through funds. The canonical advisor case: employer stock held directly + ESPP/RSU lots + a broad index fund + a target-date fund — every account-level view says "diversified," while the household owns one company four ways. Rebalance review is blind to this.
What the skill does
ROGin a US fund is Rogers Corp, not Roche); ambiguous tickers stay un-mergedPlus portability guardrails: holdings-file staleness stamps, one-level-only expansion for fund-of-funds, non-equity sleeves excluded from the equity view.
Provenance / fit
Distilled from a production family-office portfolio system I run — method and failure modes only, provider-neutral, markdown-only, styled after
thesis-tracker. A complete SKILL.md is drafted and ready. Fitsplugins/vertical-plugins/financial-analysis/skills/(shared core) or the advisors vertical alongside rebalance review.Next step
Repo forked (
sbhavani/financial-services). Happy to open the PR once placement is confirmed — or iterate on scope here first.