docs(readme): rewrite as evidence-verified project showcase

[mrnicholasbcarter-code]

Summary

Full README rewrite (306 → 264 lines), rebuilt from verified project reality rather than aspirational claims.

Removed (unverifiable or broken):

• PyPI badges and pipx install verdict-core — package returns 404, not published
• github.com/verdict/verdict-core badge links — wrong org
• "live catalog (3,900+ models)" phrasing — contradicts claims-ledger CL-007 (unsupported: catalog counts are historical observations with bounded liveness evidence)

Added / verified against the tree:

• Hero positioning from the actual repo description: Route LLM tasks by criticality. Never send prod code to a cheap model. Never burn $20/hr on formatting.
• Recommendation-engine vs. control-plane comparison table — the project's actual wedge
• 30-second demo block quoting verdict quickstart --non-interactive --dry-run output byte-for-byte (re-run and confirmed on this branch)
• Hard guarantees citing real, existing tests (test_ranker_cannot_reintroduce_excluded_candidate, test_protected_work_fails_closed_when_truth_absent, etc.)
• Full 16-command CLI surface incl. installer, setup wizard, doctor --fix, detect
• Evidence-backed public claims section linking docs/proof/EVIDENCE_INDEX.md, claims_ledger.v1.json, and docs/portfolio/PORTFOLIO_PROOF_MATRIX.md (preserves docs(portfolio): add evidence-backed conversion assets #208 intent)

Verification

• 34/34 relative links resolve; all in-page anchors valid
• 4/4 cited test names exist in tests/; test-file count corrected (73 → 81)
• 16 CLI commands and all claimed flags confirmed in verdict/cli.py
• Quickstart output matches byte-for-byte (executed on this branch)
• Pre-commit hooks pass: ruff check, ruff format, mypy --strict (90 files)
• git diff --check clean

🤖 Generated with Claude Code

[github-actions]

📊 Benchmark Results

coding

Metric	Fixed	Verdict	Delta
Success Rate	100.0%	98.5%	-1.5%
Cost/req	$0.014711	$0.006887	53.2%
p95 Latency	146ms	142ms	-5ms

reasoning

Metric	Fixed	Verdict	Delta
Success Rate	100.0%	96.5%	-3.5%
Cost/req	$0.007361	$0.004478	39.2%
p95 Latency	142ms	145ms	3ms

chat

Metric	Fixed	Verdict	Delta
Success Rate	98.0%	99.0%	1.0%
Cost/req	$0.000097	$0.001421	-1364.9%
p95 Latency	74ms	142ms	67ms