feat: standalone context-verify CLI and updated cross-platform matrix

[nathanschram]

Phase 1: Foundation (v1.6.0)

Context

ContextDocs' cross-tool compatibility table is outdated. The AI coding tool landscape has converged — most tools now support AGENTS.md, skills (SKILL.md), hooks, and custom agents. ContextDocs needs to reflect this and provide universal tooling.

Deliverables

1. bin/context-verify.sh — Standalone health scoring CLI (bash + git + jq, no plugin required)
   • 10+ automated checks: line budgets, stale paths, bridge consistency, context load, freshness
   • 0-100 score with A-F grades
   • CI mode: --ci --min-score 90 with exit codes
2. Updated docs/guides/troubleshooting.md — Replace outdated "Manual/No" compatibility matrix with March 2026 platform capabilities
3. New docs/references/platform-capabilities.md — Per-tool capability documentation for 9 platforms

Verification

• context-verify.sh produces consistent scores
• Existing tests pass: check-token-budgets.sh, validate-llms-txt.sh, test-hooks.sh

Part of the Cross-Platform ContextDocs Strategy.