docs: Fix 24 documentation audit findings

[JoshDoesIT]

Summary

Comprehensive documentation audit identified and fixed 24 findings: 7 inaccuracies, 4 inconsistencies, and 13 improvement opportunities across all project docs.

The most impactful fix: setup.sh pulled llama3 but the CLI defaulted to qwen2.5:14b, causing errors on first run for any new contributor following the docs.

Changes

• setup.sh: Default model llama3 → qwen2.5:14b to match _defaults.py
• README.md: Add html to output formats, fix tree structure, add mypy badge, document default model + resource requirements, add missing make targets
• ARCHITECTURE.md: Expand enrichment.py description, add eval_command, eval_ragas, _json_utils to supporting modules table, refine strict mode claim
• CONTRIBUTING.md: Add uv prerequisite, fix model reference, add make typecheck/check/test-all/docs, add 3 missing eval test suites, fix eval fixture schema fields
• CHANGELOG.md: Fix inaccurate "dynamic version" claim, add 6 missing features to Unreleased
• SECURITY.md: Add LLM prompt injection to scope
• CODE_OF_CONDUCT.md: Add specific contact info replacing vague "contact information available on the repository"
• mkdocs.yml: Add pymdownx.snippets extension (fixes broken docs/architecture.md), add eval API reference to nav
• docs/index.md: Add eval module to API reference list
• docs/api/eval.md: New mkdocstrings page for eval_command
• .github/SUPPORT.md: Note that Discussions requires repo enablement

Test Plan

• All existing tests pass (uv run pytest)
• New tests written following TDD (Red → Green → Refactor)
• Linting passes (uv run ruff check .)
• Type checking passes (uv run mypy src/)

Documentation-only changes — no production code modified, no new tests needed.

Checklist

• Tests written before implementation (TDD)
• Documentation updated (if applicable)
• No unused code, console logs, or dead comments
• Follows existing project coding style