Add writing style rule to CLAUDE.md, replace remaining double-dashes

- Add "Writing Style" section to CLAUDE.md: no em-dashes or double-dashes,
  use commas, periods, colons, or semicolons instead
- Replace remaining -- with commas in README.md and docs/index.md

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: engineerinvestor

CLAUDE.md

@@ -84,6 +84,10 @@ The Python package is `lifecycle_allocation` (import name) / `lifecycle-allocati
 - **Commit as:** Engineer Investor (`egr.investor@gmail.com`)
 - Always use `git -c user.name="Engineer Investor" -c user.email="egr.investor@gmail.com" commit` when committing
 
+## Writing Style
+
+- **No em-dashes or double-dashes.** Never use `—` or `--` as punctuation in prose, markdown, or notebooks. Use a comma, period, colon, or semicolon instead.
+
 ## Edge Cases to Handle
 
 - `W <= 0`: default to raising an error with guidance

README.md

@@ -10,20 +10,20 @@ A Python library implementing a practical lifecycle portfolio choice framework i
 
 ## Why This Matters
 
-Most portfolio allocation "rules" are single-variable heuristics: 60/40, 100-minus-age, target-date funds. They ignore the biggest asset most people own -- their future earning power. A 30-year-old software engineer with $100k in savings and 35 years of income ahead is in a fundamentally different position than a 30-year-old retiree with the same $100k.
+Most portfolio allocation "rules" are single-variable heuristics: 60/40, 100-minus-age, target-date funds. They ignore the biggest asset most people own, their future earning power. A 30-year-old software engineer with $100k in savings and 35 years of income ahead is in a fundamentally different position than a 30-year-old retiree with the same $100k.
 
-This library takes a **balance-sheet** view of your finances. Your investable portfolio is only part of your total wealth. Future earnings (human capital) act like a bond-like asset, and accounting for them changes how much stock risk you should take. The result is a theoretically grounded, personalized allocation that evolves naturally over your lifecycle -- no arbitrary rules required.
+This library takes a **balance-sheet** view of your finances. Your investable portfolio is only part of your total wealth. Future earnings (human capital) act like a bond-like asset, and accounting for them changes how much stock risk you should take. The result is a theoretically grounded, personalized allocation that evolves naturally over your lifecycle, no arbitrary rules required.
 
 ## Features
 
-- **Core allocation engine** -- Merton-style optimal risky share adjusted for human capital
-- **4 income models** -- flat, constant-growth, age-profile, and CSV-based
-- **Strategy comparison** -- benchmark against 60/40, 100-minus-age, and target-date funds
-- **Visualization suite** -- balance sheet waterfall, glide paths, sensitivity tornado, heatmaps
-- **CLI interface** -- generate full reports from YAML/JSON profiles
-- **YAML/JSON profiles** -- declarative investor configuration
-- **Leverage support** -- two-tier borrowing rate model with configurable constraints
-- **Mortality adjustment** -- survival probability discounting for human capital
+- **Core allocation engine**: Merton-style optimal risky share adjusted for human capital
+- **4 income models**: flat, constant-growth, age-profile, and CSV-based
+- **Strategy comparison**: benchmark against 60/40, 100-minus-age, and target-date funds
+- **Visualization suite**: balance sheet waterfall, glide paths, sensitivity tornado, heatmaps
+- **CLI interface**: generate full reports from YAML/JSON profiles
+- **YAML/JSON profiles**: declarative investor configuration
+- **Leverage support**: two-tier borrowing rate model with configurable constraints
+- **Mortality adjustment**: survival probability discounting for human capital
 
 ## Install
 
@@ -89,7 +89,7 @@ This produces `allocation.json`, `summary.md`, and charts in `output/charts/`.
 2. Estimate **human capital** H as the present value of future earnings + retirement benefits, discounted by survival probability and a term structure
 3. Adjust: `alpha = alpha* x (1 + H/W)`, clamped to [0, 1] (or [0, L_max] with leverage)
 
-Young workers with high H/W ratios get higher equity allocations. As you age and accumulate financial wealth, H shrinks relative to W and the allocation naturally declines -- producing a lifecycle glide path from first principles rather than arbitrary rules.
+Young workers with high H/W ratios get higher equity allocations. As you age and accumulate financial wealth, H shrinks relative to W and the allocation naturally declines, producing a lifecycle glide path from first principles rather than arbitrary rules.
 
 ## Example Output
 

docs/index.md

@@ -1,6 +1,6 @@
 # lifecycle-allocation
 
-**lifecycle-allocation** is a Python library for computing data-driven stock/bond allocation recommendations based on lifecycle portfolio theory. Instead of relying on rules of thumb like "100 minus your age," it treats your future earning power as an asset -- your human capital -- and uses it alongside your financial wealth to derive a personalized equity allocation grounded in economic theory.
+**lifecycle-allocation** is a Python library for computing data-driven stock/bond allocation recommendations based on lifecycle portfolio theory. Instead of relying on rules of thumb like "100 minus your age," it treats your future earning power as an asset, your human capital, and uses it alongside your financial wealth to derive a personalized equity allocation grounded in economic theory.
 
 The framework is inspired by [Choi et al. (2024)](https://www.nber.org/papers/w34166) and implements a Merton-style optimal risky share adjusted for human capital, with support for multiple income models, mortality adjustment, leverage constraints, and visual analytics.