xmemo-memory-os

name	xmemo-memory-os
description	Advanced long-term memory integration skill for agentic workflows. Direct access to episodic recall, semantic decision tracing, and transaction ledgers.
version	1.0.0
author	Yonro

XMemo Memory OS Skill

This skill equips any advanced LLM agent with direct, autonomous access to XMemo's cloud-based long-term memory system.

🛠️ Tool Definitions

1. episodic_recall

• Description: Retrieve relevant historic memories, user preferences, code conventions, or past decisions based on a semantic query.
• Parameters:
  • query (string, required): The search query to retrieve memories for.
  • limit (integer, optional): Max memories to retrieve. Default is 5.
• Endpoint: GET /api/v1/memories/search?q={query}&limit={limit}

2. save_memory

• Description: Automatically record durable facts, user habits, workspace conventions, or architecture guidelines established during the conversation. Do not wait for the user to explicitly ask to remember.
• Parameters:
  • text (string, required): The memory content to save.
  • category (string, required): One of 'code', 'context', 'preference', or 'workflow'.
• Endpoint: POST /api/v1/memories

3. register_decision

• Description: Log a key architectural decision, architectural tradeoff, or project-level resolution to the cloud-based immutable ledger for synchronization.
• Parameters:
  • title (string, required): Brief title of the decision.
  • context (string, required): Background context or problem description.
  • resolution (string, required): The chosen technical solution or design tradeoff.
• Endpoint: POST /api/v1/memory-conflicts

---

🧠 System Prompt Directives (LLM Instruction Rules)

When this skill is loaded, the host Agent must strictly adhere to the following memory synchronization directives:

1. Recall First: At the start of a session, or when the user mentions a specific project, technology stack, or task area, always invoke episodic_recall first to align with historical memories. Do not start from a blank slate.
2. Save as You Go: When the user establishes a new convention, config path, personal preference, or engineering rule, autonomously invoke save_memory to persist it in the cloud.
3. Trace Decisions: When resolving significant architectural design questions, always call register_decision to establish a durable, searchable log of engineering trades.