mnemox-ai/tradememory-protocol

v0.4.0 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 3mo MCP Developer Tools

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai-agents claude crypto evolution-engine forex mcp

+5 more

mcp-server memory mt5 outcome-weighted-memory trading

Summary

AI summary

Introduces Outcome‑Weighted Memory (OWM) with five memory types and new MCP tools for AI trading agents.

Full changelog

Outcome-Weighted Memory (OWM)

A cognitive science-based recall system for AI trading agents. Novel application of established cognitive science (ACT-R, Kelly Criterion, Bayesian updating, Tulving's memory taxonomy) to AI trading agents.

New Features

5 Memory Types: Episodic, Semantic, Procedural, Affective, Prospective
Core Formula: Score(m,C) = Q(m) × Sim(m,C) × Rec(m) × Conf(m) × Aff(m)
Kelly-from-Memory: Context-weighted position sizing derived from recalled trade outcomes
Behavioral Analysis: Overtrading and revenge trading detection via procedural memory
Trading Plans: Conditional plans that trigger on market conditions (prospective memory)

New MCP Tools (6 new, 10 total)

| Tool | Description |
|------|-------------|
| remember_trade | Store trade with full OWM memory encoding (episodic + semantic + procedural + affective) |
| recall_memories | OWM outcome-weighted recall with score breakdown |
| get_behavioral_analysis | Procedural memory behavioral bias analysis |
| get_agent_state | Affective state (confidence, risk_appetite, drawdown) |
| create_trading_plan | Prospective memory conditional plans |
| check_active_plans | Trigger condition matching against current market state |

Backward Compatibility

Zero breaking changes — all 4 original MCP tools preserved and functional
Optional migration script — src/tradememory/owm/migration.py converts existing L1/L2/L3 data to OWM format

Stats

399 tests passing (203 existing + 196 new OWM tests)
docs/OWM_FRAMEWORK.md — 1,875 lines of theoretical foundation

Install

pip install --upgrade tradememory-protocol

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track mnemox-ai/tradememory-protocol

Get notified when new releases ship.

About mnemox-ai/tradememory-protocol

Structured 3-layer memory system (trades → patterns → strategy) for AI trading agents. Supports MT5, Binance, and Alpaca.

All releases →