agent-memory
/agent-memory
Use when implementing or managing persistent, hierarchical memory systems for AI agents. Covers cross-session state, fact supersession, and self-managed memory tools to enable long-term recall and adaptive agent behavior.
- Author
- VoDaiLocz ↗
- Repository
- kilo-kit-mcp ↗
- Licence
- Apache-2.0
- Last repository change
- 24.10.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 5
No signals detected. Repository changed on 24.10.2026, without references to retired models or APIs.
Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.
What it does and when to use it
Use when implementing or managing persistent, hierarchical memory systems for AI agents. Covers cross-session state, fact supersession, and self-managed memory tools to enable long-term recall and adaptive agent behavior.
Original author description
Use when implementing or managing persistent, hierarchical memory systems for AI agents. Covers cross-session state, fact supersession, and self-managed memory tools to enable long-term recall and adaptive agent behavior.
How it works
- Overview
- When To Use
- Core Concepts
- Memory Architecture
- Implementation Patterns
- CRUD Operations
- Context Window Management
- Quality Gates
- References
SKILL.md preview
--- name: "agent-memory" description: >- Use when implementing or managing persistent, hierarchical memory systems for AI agents. Covers cross-session state, fact supersession, and self-managed memory tools to enable long-term recall and adaptive agent behavior. --- # Agent Memory Framework ## Overview The `agent-memory` skill provides a standardized architectural approach to building intelligent memory systems for agents within the KILO-KIT ecosystem. It bridges the gap between ephemeral context windows and durable, long-term storage, enabling agents to maintain user preferences, project-specific conventions, and debugging history across multiple sessions. ## When To Use - When designing systems that must persist state across independent interaction sessions. - When an agent needs to manage large volumes of user-specific facts that exceed the context window. - When implementing self-managed memory tools (MemGPT/Letta patterns) to allow agents to control their own knowledge base. - When building systems requiring automatic entity updates (Mem0 pattern) to resolve conflicting or stale information. ## Core Concepts - **Memory Hierarchy**: Differentiating between transient working context and persistent knowledge. - **Fact Extraction**: Identifying core entities, preferences, and relationships from conversational flow. - **Supersession**: Automatically replacing outdated facts with new information to maintain "ground truth." - **Temporal Validity**: Tracking the lifespan and relevance of memory entries over time. - **Persistence**: Ensuring data survives agent resets or session termination. ## Memory Architecture The framework defines four distinct tiers of memory: 1. **Working Context (RAM)**: The immediate token window. Ephemeral, high-speed, and limited in capacity. 2. **Episodic Memory**: Logged history of past interactions, enabling agents to query "what we discussed last time." 3. **Semantic Vector Store**: Long-term storage for semantic concepts, documentation snippets, and project conventions, retrieved via similarity search. 4. **Archival Storage (Disk)**: Cold storage for large documents or historical artifacts that are rarely needed but must be maintained. ## Implementation Patterns ### Fact Extraction & Supersession (Mem0 Pattern) - Implement extraction loops that analyze messages for key-value pairs (e.g., `user_preference: dark_mode`). - When a new fact conflicts with an old one, perform an "update" (supersession) rather than appending duplicates. This ensures the agent always acts on the most recent truth. ### Temporal Validity Windows (Zep/Graphiti Pattern) - Annotate memory items with timestamps and `TTL` (Time-To-Live). - Implement background cleanup processes to purge or archive expired or invalidated information based on these windows. ### Self-Managed Memory Tools (Letta/MemGPT Pattern) - Equip agents with dedicated function calls: - `memory_write(key, value, importance)` - `memory_read(query)` - `memory_archive(id)` - Empower the agent to decide when to offload items from the working context to memory storage. ## CRUD Operations - **Create**: Automatically extract new knowledge from successful interactions and persist to storage. - **Read**: Perform semantic search across the vector store and episodic history to augment the current context. - **Update**: Supersede stale facts with verified new information to prevent hallucinations or conflicting behavior. - **Delete/Archive**: Offload deprecated or low-value information to archival storage to maintain high-signal density in memory. ## Context Window Management Use the following heuristic for resource allocation: 1. **Critical Path**: Keep in active context if needed for the current turn. 2. **Supportive Memory**: Fetch from vector store if the task requires historical context. 3. **Episodic Recall**: Search past logs only when explicitly prompted or when a long-term pattern is requested. 4. **Avoid Bloat**: Do not put long documents directly into the prompt; use `read_resource` or semantic retrieval instead. ## Quality Gates - **Consistency Check**: Run automated verification to detect conflicting memory entries. - **Persistence Test**: Verify that user-defined preferences are available after agent restart. - **Signal-to-Noise Ratio**: Periodically audit the memory store to remove redundant, outdated, or low-relevance facts. - **Latency Monitoring**: Ensure that retrieving from memory does not introduce unacceptable delays in response time. ## References - [Mem0 Documentation](https://mem0.ai) - [Letta / MemGPT Architecture](https://letta.com) - [Zep Memory Framework](https://getzep.com) - [KILO-KIT Internal Documentation](https://kilo-kit.io) - [Graphiti Memory Patterns](https://graphiti.dev)
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · VoDaiLocz/kilo-kit-mcp | 90 | Apache-2.0 |
| TerminalSkills/skills | 100 | Apache-2.0 |
| komluk/scaffolding | 100 | MIT |
| nirholas/three.ws | 100 | Apache-2.0 |
| nimadorostkar/Claude-Skills-collection | 100 | MIT |
| alirezarezvani/claude-skills | 100 | MIT |
| norifumi3/cc-marketplace | 100 | MIT |
| OpenHands/extensions | 92 | MIT |
| HelixDevelopment/agent | 92 | MIT |
Database snapshot: 13.11.2026 · package version a9508f5d3e49 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 90conversation-memory
Persistent memory systems for LLM conversations including short-term, long-term, and entity-based memory Use when: conversation memory, remember, memory persistence, long-term memory, chat history. - 90conversation-memory
Persistent memory systems for LLM conversations including short-term, long-term, and entity-based memory Use when: conversation memory, remember, memory persistence, long-term memory, chat history. - 90conversation-memory
Persistent memory systems for LLM conversations including short-term, long-term, and entity-based memory Use when: conversation memory, remember, memory persistence, long-term memory, chat history. - 100simplemem
Build persistent LLM agent memory with SimpleMem - semantic lossless compression with 30x token reduction. MCP server (cloud/self-hosted), Python API, cross-session memory. Use when: adding long-term memory to agents, cross-session context recall, semantic dialogue search, memory consolidation. - 90mem0-memory
Use optional Mem0 memory tools for focused cross-session recall and short verified durable facts. Use when the user asks to remember or forget something, when resuming work that may need prior context, or after verified progress creates a stable fact worth retaining. - 100memory-systems
Design and implement memory architectures for agent systems. Use when building agents that need to persist state across sessions, maintain entity consistency, or reason over structured knowledge.