eval-viewer: .md output files displayed as raw text instead of rendered markdown

The eval viewer in `skill-creator` displays `.md` output files as raw text
inside `

` tags. Since most skills produce markdown outputs with tables,

headings, code blocks, and lists, the review experience is significantly

degraded — reviewers see raw `#`, `|`, and triple-backtick syntax instead

of formatted content.


**Steps to reproduce:**

1. Run `/skill-creator` evals on any skill that outputs `.md` files

2. Open the generated review HTML (via `generate_review.py`)

3. Observe that `.md` files show raw markdown syntax

**Expected:** `.md` files rendered as formatted HTML (headings, tables, code blocks)

**Actual:** Raw markdown text in monospace `
` block


I have a small patch ready (~54 lines, 1 file) that adds markdown rendering

via marked.js with graceful fallback. Happy to open a PR if this approach

seems reasonable.