trader-mcp/TODO.md

Trader-MCP TODO

Trader-MCP concerns only strategy behavior and the surfaces strategies must expose.

Strategy taxonomy

• Reduce strategy set to a small, orthogonal registry.
• Treat parameter variants as modes inside a strategy, not separate strategies.
• Keep at minimum: idle, trend_following, mean_reversion, breakout, grid, event_driven, defensive.
• Remove or fold redundant strategy names that differ only by tuning.

Strategy metadata

• Add explicit expects / avoids regime metadata to every strategy.
• Declare natural habitat fields for trend, volatility, event risk, and liquidity.
• Add a risk profile per strategy.

Example: strategy capability declaration

{
  "name": "mean_reversion",
  "capabilities": ["fade_extremes", "range_entry", "range_exit"],
  "expects": {
    "trend": "none",
    "volatility": "low",
    "event_risk": "low",
    "liquidity": "normal"
  },
  "avoids": {
    "trend": "strong",
    "volatility": "expanding",
    "event_risk": "high"
  },
  "risk_profile": "medium"
}

Example: strategy contract to Hermes

{
  "name": "trend_following",
  "inputs": ["market_regime", "sentiment_pressure", "liquidity_state"],
  "outputs": ["target_position", "risk_mode", "reason"],
  "exposes": ["status", "recent_changes", "decision_history"],
  "confidence_threshold": 0.65,
  "minimum_hold_minutes": 15
}

Strategy surface

• Expose clean strategy state.
• Provide strategy status, recent changes, and decision history.

Execution feedback

• Feed fills, slippage, execution quality, and stress into strategy selection.
• Let execution degradation influence strategy selection.

Cleanup / simplification

• Separate signal-based strategies from structure-based ones, especially mean_reversion vs grid.
• Avoid adding micro-strategies until the minimal taxonomy is stable.
• Keep strategy names human-readable and stable.