[Repo Assist] Fix ChiSquared.InvCDF scale parameter bug; add InvCDF tests

[github-actions]

🤖 This is an automated PR from Repo Assist.

Problem

ChiSquared.InvCDF returned drastically wrong quantiles. For example, ChiSquared.InvCDF 20. 0.1 returned ≈ 3.11 instead of the correct ≈ 12.44.

Root cause

ChiSquared.InvCDF delegates to Gamma.InvCDF, but was passing beta = 1/2 while FSharp.Stats.Distributions.Continuous.Gamma uses the scale parameterisation (Mean = α × scale). The chi-squared distribution corresponds to Gamma(α = dof/2, scale = 2), so the correct value is beta = 2.0, not 0.5.

// Before (wrong — β = rate, but Gamma uses scale)
let beta = 1. / 2.0
// After (correct)
let beta = 2.0
```

The bug was already noted in the title of open PR #363 ("fix ChiSquared.InvCDF scale bug"), which implements InvCDF for several other distributions. This PR provides the minimal fix and a dedicated test suite so the bug cannot regress regardless of how #363 evolves.

## Changes

| File | Change |
|------|--------|
| `src/.../ChiSquared.fs` | Fix `beta = 1/2 → 2.0` in `InvCDF`; add explanatory comment |
| `tests/.../DistributionsContinuous.fs` | Add `testList "ChiSquared.InvCDF tests"` with 7 cases |

### New tests

- `p = 0` → 0 (boundary)
- `p = 1` → +∞ (boundary)
- Round-trip: `CDF(InvCDF(p)) ≈ p` at p = 0.5, 0.95, 0.01
- Known values derived from the CDF reference table already in the file (Williams 1984)

## Test Status

```
dotnet test  →  Passed!  Failed: 0, Passed: 1201, Total: 1201

All 1201 existing tests continue to pass; 7 new tests added for ChiSquared.InvCDF.

Closes part of #262 (InvCDF gaps).

Generated by 🌈 Repo Assist, see workflow run. Learn more.

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/repo-assist.md@7ee2b60744abf71b985bead4599640f165edcd93