Agent memory

Type: types/tag-readme.md

This tag gathers what agents retain across sessions and how that retention becomes useful: persistence and storage, recall and activation, extracting memory from traces, authority over what is kept, lifecycle operations such as retiring and superseding, and comparisons of external agent memory systems. The anchor is Designing a Memory System for LLM-Based Agents, the synthesis from which the memory-system requirements are derived. Members span kb/notes/, the external memory-system comparisons in kb/agent-memory-systems/, and system analyses in kb/agentic-systems/. Nearby but different: context-engineering covers getting knowledge into any bounded call; a note belongs here when its question is what is retained between sessions and under what authority, not only how a call is assembled. This head is selective; use a scoped tag search for full membership.

Core Frame

• Designing a Memory System for LLM-Based Agents — the anchor synthesis: derives the memory-system requirements from bounded context, consumer needs, artifact governance, and retrieval's limits
• Agent memory needs discoverable, composable, trusted knowledge under bounded context — minimal artifact-quality frame for remembered knowledge that must improve future action
• Agent memory is a crosscutting concern, not a separable niche — decomposes memory into storage, retrieval/activation, and learning rather than treating it as one pluggable subsystem
• Raw accumulation does not create usable memory — separates preserving material from making it searchable, scoped, trustworthy, and lifecycle-managed
• Memory design adds operational axes to artifact analysis — extends artifact analysis with capture, derivation, activation, authority assignment, lifecycle, and evaluation policies

Requirements Map

• Agent Memory Requirements — navigation hub for concrete requirements extracted from the memory-system design synthesis
• Activate Behavior-Changing Memory Before The Mistake — memory must reach the operative context before the relevant action, not only after explicit search
• Make Authority Explicit — governance rule for who can read, write, promote, activate, enforce, revise, and retire memory
• Keep Lineage And Compiled Views From Drifting — generated cues, prompt files, indexes, and assistant-specific views need provenance and staleness checks
• Evaluate Memory By Effects, Not By Existence — memory succeeds when downstream tasks, behavior, context efficiency, or lineage improve
• Retire, Redact, Supersede, And Relax Memory — lifecycle operations for changing, invalid, sensitive, or over-constrained memory

Boundaries And Failure Modes

• Active work state is not retrospective memory or chat history — separates live task state from retained retrospective memory
• A context-operation interface bounds the projections its policy can realize — separates retained substrate, available projection operations, controller policy, and active-context exposure
• Preserve Evidence Without Making History The Next Context — keeps trace evidence available for audit and extraction without loading raw history by default
• Flat memory predicts specific cross-contamination failures that are empirically testable — predicts search pollution, identity scatter, and insight trapping when memory roles collapse
• Trace-extracted memory earns authority per operation, not at capture — trace-extracted records become knowledge only after operations such as verification, abstraction, and consultation
• Bottom-up structure inference needs capture at the decision surface, not the state — relation inference works only when capture preserves the decision-shaped "why"

External Checks

• The adaptation survey corroborates memory requirements but misses artifact governance — compares external agentic-adaptation taxonomy against the memory requirements map
• Memory management policy is learnable but oracle-dependent — AgeMem shows memory policy can be learned when task-completion oracles exist
• Three-space agent memory echoes Tulving's taxonomy but the analogy may be decorative — tests whether knowledge, self, and operational memory earn separate lifecycle treatment

Memory-System Comparisons

• Agentic memory systems: comparative review — cross-system findings from the code-grounded reviews: storage choices, read-back, trace learning, and how rarely lifecycle curation appears
• Systems table — one row per reviewed memory system across storage, read-back, targeting, trace learning, and enforcement
• Review framework design — why the comparison matrix has the columns it has
• The full catalogue of reviewed systems is the agent-memory-systems index

Related Tags

• Learning theory — memory only matters when retained artifacts change future behavior
• Context engineering — the wider area: retained memory becomes useful only through selection and activation into bounded context; most requirements notes carry both tags
• Computational model — agent memory is loaded through bounded calls and scheduling decisions

Other tagged notes

• A checked outcome licenses retaining an episode, not abstracting its explanation - One result-only check can warrant retaining an episode as evidence, but abstracting its explanation also needs evidence about a faithful producing process and an explicit scope boundary
• Adaptation signals choose pressure; artifact analysis chooses the retained surface - Maps agentic-adaptation signals onto artifact-analysis axes so KB learning records which retained surface changes, what authority it gains, and how to review it
• AI Agents in Depth - Whole-book comparison of AI Agents in Depth with Commonplace, separating broad architectural convergence from differences in memory admission, epistemic warrant, governance, and orchestration
• An accepted edit verifies the change, not the rule - Human acceptance of an edit is a strong oracle for 'this change was wanted here' but a weak oracle for 'this generalizes' — mining rules from accepted edits inherits instance-level verification while the generalization step stays oracle-poor
• beads_rust - beads_rust as a local active-work and coordination substrate: transactional CLI claims and workflow gates, explicit external execution and Git boundaries, and weaker parity across MCP, inherited-context, and shipped instruction paths
• Create Memory Directly - Direct memory creation preserves live understanding by writing useful artifacts before later trace extraction loses structure
• Explicit retention provides direct targets for selective revision - Explicit artifacts give a learner direct targets for inspecting and revising commitments; durability, writability, and effective addressability still depend on the boundary and available operations
• Import External Knowledge Into Internal Form - Agent memory systems need import paths when authoritative project knowledge already exists outside the memory substrate
• Memory-backed personalization can look like model improvement - Distinguishes user-specific gains supplied by retained intent from gains in the model that interprets the assembled context.
• Open-domain memory retention needs a declared output spec - Explains why an input stream alone can't answer 'what to store' in open-domain memory design; a declared output spec supplies the missing inclusion criterion.
• Parametric reproduction alone cannot replace an authoritative record - Reproducing a record's content does not transfer its authority. Replacement requires a governed artifact with stable identity, integrity, contestability, and attribution; mutable records also require currentness and addressable revision.
• Promote Only When Future Value Exceeds Maintenance Cost - Candidate memory should become durable only when future retrieval or activation value exceeds review and maintenance cost
• Promotion selects for unreliable activation, and the regress ends only at an external trigger - Recasts promotion from 'the consumer lacks this' to 'the consumer will not apply this unprompted', and requires delivery to have a root firing event independent of that prior activation
• Retaining episode evidence keeps a distilled rule open to re-examination - Keeping relevant episode evidence and its relation to a distilled rule preserves a route for re-examining that rule; reconstruction, comparative value, and correct generalization still require testing
• Rule-based context selection needs a pre-existing signal - A rule-based selector can target one case only when a rule-ready signal already distinguishes it; otherwise the system must wait, load broadly, or infer relevance from task and candidate content
• Serve Multiple Consumers, Not One Retrieval Interface - Memory systems need multiple surfaces because acting, scheduling, review, learning, governance, and active work consume memory differently
• Use Trace Extraction As Meta-Learning - Trace extraction is an after-the-fact learning path that must respect signal quality, review, and readable-artifact versus distributed-parametric learning boundaries