feat: relevance ranking, evals, chronicle, embeddings, skill drift-check

[BarrettHolien]

Summary

Five additive features that graft hybrid-search ranking, eval scaffolding, and history tooling onto graphify's structural spine. Each is a self-contained module + CLI command + tests; no existing behavior changes unless the new flags are used.

1. Ranking (graphify/ranking.py) — ranks query results by Reciprocal Rank Fusion over lexical + structural (graph proximity, hub centrality, community) + optional semantic backends, replacing degree-only render order. New query flags --top-k, --explain, --semantic; also exposed on the MCP query_graph tool via a shared core (serve._resolve_and_rank / rank_query_nodes).

2. Evals (graphify/evals.py) — graphify bench scores P@k / recall / MRR / nDCG against a JSONL fixture through the real ranking pipeline. --init scaffolds a fixture; --save / --replay gate regressions (exit 1) for CI.

3. Embeddings (graphify/embed.py) — opt-in graphify embed builds a local sidecar (Ollama nomic-embed-text with search_query:/search_document: prefixes, or sentence-transformers; no API calls). --semantic fuses cosine similarity into ranking and seeds traversal when lexical matching finds nothing.

4. Chronicle (graphify/chronicle.py) — graphify chronicle OLD NEW (or --rev via git show) diffs two graph snapshots: node/edge deltas, god nodes emerged/vanished, community shifts. API-free.

5. Skill drift (graphify/skill_migrations.py) — graphify skill status|check-update reports installed-skill-vs-package drift (exit 1 for CI) with a migration registry.

Skill fragments document the new query flags plus a new analysis.md reference; all skillgen artifacts regenerated and blessed (skillgen --check / --audit-coverage pass).

Testing

• 3,161 passed, 31 skipped on this branch rebased onto current v8 (4c075f9). The only failures in my environment (8 in test_llm_backends/test_ollama*, missing optional openai module) reproduce identically on clean v8, so they're unrelated to this change.
• New test files: test_ranking.py, test_evals.py, test_embed.py, test_chronicle.py, test_skill_migrations.py.

Notes for maintainers

• The commit bumps version 0.9.12 → 0.10.0 and adds a CHANGELOG entry — happy to drop those if you prefer to manage releases yourselves.
• This is intentionally one commit of five cohesive features (ranking is the base the others build on), but I'm glad to split it into smaller stacked PRs if that's easier to review.

[BarrettHolien]

Note: this supersedes #1760 (same change set, rebased onto current v8 and re-tested).