Insurance Policy Decoder Skill

Insurance policies are read twice: at signing (by no one) and after the loss (too late). This skill does the first reading properly — what each coverage line pays in a real scenario, which exclusions swallow which promises, and whether "covered" means replaced-new or depreciated-to-pennies. The declarations page is marketing; the exclusions and definitions sections are the policy.

What This Skill Produces

• A coverage-by-coverage decode with a concrete payout scenario for each ("kitchen fire, $40k damage → policy pays X because…")
• Ranked red flags: exclusions that gut headline coverage, sublimits, ACV traps, coinsurance penalties
• The ACV vs. replacement-cost math, shown on the user's actual property
• Questions for the agent — gaps found, endorsements worth pricing, and what to get in writing

Required Inputs

Ask for these only if they aren't already provided:

• The policy documents — declarations page at minimum; exclusions/definitions sections if available. With only a declarations page, decode what's visible and list the sections still needed — the exclusions are where the reading matters.
• What's being protected — home value and contents ballpark, or vehicle + how it's used; anything unusual (home business, expensive equipment, a finished basement in a rain-prone area).
• Their worry list — the losses they actually fear; the decode ranks against those.

Framework: Severity Scale

• 🔴 Can cost you real money — actual-cash-value settlement on roof/contents (depreciation eats the payout), water/flood/sewer-backup exclusions (the most commonly discovered-too-late gap — flood is almost never in a standard policy), sublimits far below stated property ("jewelry: $1,500 total"), coinsurance clauses (insure below X% of value → every claim penalized proportionally), ordinance-of-law gaps (rebuilding to current code isn't covered by default), business-use exclusions voiding claims for home-office equipment or delivery driving.
• 🟡 Unusual — probe before renewal — high or percentage-based wind/hail deductibles (2% of dwelling ≠ 2% of the claim), depreciation on partial roof claims, named-perils contents coverage masquerading as comprehensive, rental-car and loss-of-use caps that run out mid-repair.