v2.0.0 — Memory Layer, Audit Trail, Cost Tracking

⚡ Maestro v2.0.0

🆕 Memory Layer

Maestro now remembers what happened across sessions:

.maestro/
├── context.md         ← project context (replaces .maestro.md)
├── decisions.jsonl    ← append-only decision log
├── audit.jsonl        ← every command with cost + duration
└── sessions/
    └── 2026-04-26_fix_auth.md

• Backward compatible — .maestro.md users change nothing
• Opt-in — .maestro/ is created only when you use /capture

🆕 3 New Commands (22 → 25)

Command	Purpose
/capture	Save session summary — persist decisions and next steps
/reflect	Effectiveness scorecard — which skills work, which fail
/recap	Quick last-session recovery

🆕 Cost Estimation

Approximate cost tracking for Claude, GPT-4, Gemini, o3, o4-mini. Accuracy: ±20%.

🆕 3 New MCP Tools (7 → 10)

• maestro_write_decision — append decisions via MCP
• maestro_read_decisions — read decision history
• maestro_read_audit — read audit trail with cost

🔧 Pre-v2 Fixes

• P0: npm publish blocker fixed (esbuild + inline @maestro/core)
• P2: Extension bundle 5.45MB → 124KB (replaced js-tiktoken with heuristic)
• Testing: 37 unit tests in @maestro/core (vitest)

📦 Bundle Sizes

Package	Size
Extension	124KB
MCP Server	136KB
VSIX	375KB

Full Changelog

See CHANGELOG.md