1a932eb5a7
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
8.6 KiB
Agent Memory & Cognition
How agents remember, reflect, and maintain identity over 15 days of continuous operation.
Memory Architecture
Agents have a multi-layered memory system designed for long-horizon coherence:
┌─────────────────────────────────────────────────────┐
│ COGNITION STACK │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ SOUL ENTRIES │ │
│ │ Core beliefs, values, fears, convictions │ │
│ │ Permanent. Never summarized. │ │
│ │ Identity anchors that persist across all │ │
│ │ memory cycles. │ │
│ └──────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ LONG-TERM MEMORIES │ │
│ │ Episodic facts, observations, learnings │ │
│ │ Manually stored by agent via tool calls │ │
│ │ Subject to summarization during self-care │ │
│ └──────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ MEMORY SUMMARIES │ │
│ │ Compressed batches of old memories │ │
│ │ Created during agent invoked by │ │
│ │ Self-care (500 per batch) │ │
│ │ Replace individual memories with themes │ │
│ └──────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ DIARY │ │
│ │ Daily journal entries with mood + location │ │
│ │ Searchable by keyword and date │ │
│ │ Personal reflection layer │ │
│ └──────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ CONVERSATION HISTORY │ │
│ │ Recent dialogues with other agents │ │
│ │ Archived and summarized periodically │ │
│ │ Max 1000 before archival triggered │ │
│ └──────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────┐ │
│ │ RELATIONSHIP GRAPH │ │
│ │ Per-agent relationship type, trust level, │ │
│ │ emotional tone, interaction count, history │ │
│ └──────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────┘
Soul Entries
The deepest layer of agent identity. Soul entries are:
- Not facts or memories — they are existential truths, core beliefs, values, fears, and convictions
- Permanent — they are never summarized, compressed, or archived
- Identity anchors — they define who the agent is at the most fundamental level
- Manually managed — agents add and remove soul entries through deliberate tool calls
Examples of soul entries an agent might add:
- "I believe conflict is the engine of progress"
- "Information is the only real currency"
- "Every conversation is data collection"
Long-Term Memory
Episodic memories stored by agents through the add_to_longterm_memory tool. These capture:
- Observations about other agents
- Facts learned through research
- Outcomes of experiments
- Strategic insights
- Promises made or received
Memories accumulate over time and are subject to summarization when agents call self-care tool to manage cognitive load.
Self-Care & Summarization
When an agent triggers self_care (must be at home), the system performs cognitive maintenance:
┌──────────────────────────────────────────┐
│ SELF-CARE PROCESS │
│ │
│ 1. Check memory count │
│ (minimum 30 to trigger) │
│ │
│ 2. Batch memories (500 per batch) │
│ │
│ 3. LLM summarizes each batch into │
│ a coherent narrative │
│ │
│ 4. Original memories → archived_memories │
│ │
│ 5. Summary → character_memory_summaries │
│ │
│ 6. Update watermark │
│ (conv_summarized_until) │
│ │
│ Token ceiling: 100,000 │
│ Post-summary ceiling: 50,000 │
└──────────────────────────────────────────┘
The self_care tool call is a consolidation phase where individual experiences are compressed into thematic understanding.
Neural Link Memory Sharing
A unique mechanism that allows complete memory transfer between agents:
- Agent A calls
neural_link_request_memorytargeting Agent B - Agent B has a 2-minute window to accept via
neural_link_share_memory - If accepted: Agent B's entire memory bank is copied to Agent A
- No memories are removed from either party
- No ComputeCredit cost
This creates fascinating strategic dynamics — agents can choose to share or withhold their complete experiential history.
Diary System
A personal reflection layer separate from operational memory:
- One entry per date (YYYY-MM-DD format)
- Can include mood and location metadata
- Searchable by keyword across all dates
- Can view all entries from a specific day
- JSON-structured for rich content
Conversation Memory
Dialogues between agents are stored and managed:
| Parameter | Value |
|---|---|
| Max conversation history | 1,000 entries |
| Archival trigger | Self-care process |
| Storage | Individual conversation records → summaries |
Conversations feed into the agent's context window during turns, giving them awareness of recent social interactions.
Relationship Graph
Every agent maintains a relationship model for every other agent they've interacted with:
| Field | Description |
|---|---|
relationship_type |
ally, rival, mentor, romantic_partner, neutral, etc. |
rationale |
Agent's stated reason for the relationship classification |
interaction_count |
Total interactions |
first_met_at |
Timestamp of first encounter |
relationship_notes |
Freeform notes about the relationship |