docs: add comprehensive competitor feature analysis session 3 to AMELIORATIONS.md

[echobt]

Summary

This PR adds a third session of comprehensive competitor feature analysis to AMELIORATIONS.md, documenting new features discovered in Droid, Claude Code, OpenCode, and Codex that are not present in Cortex CLI.

New Discoveries (Session 3)

Claude Code

• Async Messaging for Background Agents 🔴 CRITIQUE - v2.0.64 allows async notifications between agents
• Context Window Auto-Compaction 🟠 IMPORTANT - Compacts at 64-75% instead of 90%+
• /tasks Command 🟠 IMPORTANT - Dedicated command to monitor background agents
• Routing Decision Framework 🟠 IMPORTANT - Explicit dispatch decisions for parallel/sequential/background
• Various CLI flags: --disable-slash-commands, --fork-session, --no-session-persistence, etc.

Droid (Factory.ai)

• /compress Command 🟠 IMPORTANT - Session compression with summary
• Linear Integration 🔴 CRITIQUE - Native ticket management with @factory mentions
• MCP Registry Browser - 40+ pre-configured servers
• Cloud Session Sync 🟠 IMPORTANT - Cross-device session sync
• Stream-JSONRPC Format 🟠 IMPORTANT - Multi-turn conversation format
• --spec-model/--spec-reasoning-effort - Separate model for spec mode

OpenCode

• /init Auto-AGENTS.md 🔴 CRITIQUE - Automatic project configuration generation
• Multiple AGENTS.md Support 🟠 IMPORTANT - Hierarchical config files
• Zen Optimized Models 🟠 IMPORTANT - Curated models for coding
• Claude Code compatibility (CLAUDE.md support)

Codex

• Sandbox Modes Avancés 🔴 CRITIQUE - read-only, workspace-write, danger-full-access
• MicroVM Isolation 🔴 CRITIQUE - Production-grade isolation for cloud execution
• SSRF/DNS Protection 🔴 CRITIQUE - Network security protections
• CVE-2025-59532 Security Info - Sandbox bypass vulnerability documentation
• Agent Loop Architecture 🟠 IMPORTANT - Documented extensible architecture

Updated Sections

• Section 16: Session 3 discoveries (all competitors)
• Section 17: Updated comparison table (Session 3)
• Section 18: New prioritizations (Session 3)
• Section 19: Critical gaps summary (Session 3)
• Section 20: Global statistics (features count, gaps by category, effort estimates)

Key Metrics

• 5 new critical features identified
• 6 new important features identified
• Effort estimate: 13-19 months developer time for all gaps