scoped_memory
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SCOPED_MEMORY_HOME | Yes | The absolute path to the directory where scoped memory data is stored. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| memory_init_projectA | Create or reuse a stable project identity in .scoped-memory/project.json. |
| memory_rememberB | Append a user, project, or session memory. Reusing a topic supersedes it without rewriting history. |
| memory_recallC | Build a deterministic token-budgeted context packet. Other projects are excluded unless their IDs are explicitly inherited. |
| memory_checkpointC | Store a structured task checkpoint for later continuation. Use concise model-readable facts, not conversation prose. |
| memory_ingest_projectA | Scan a project deterministically into compact EIR/1 facts and dependency edges; no model summary is used. |
| memory_engineering_contextB | Read a bounded EIR/1 engineering packet. Returns machine-readable facts without natural-language translation. |
| memory_forgetB | Append a tombstone for a topic; prior history remains auditable. |
| memory_statusB | Show store paths, counts, current project identity, and append-log integrity. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 8 tools
Each tool has a distinct responsibility: ingestion, recall, forgetting, status, checkpointing, and project init. However, memory_recall and memory_engineering_context both produce context packets, and memory_remember overlaps somewhat with memory_ingest_project and memory_checkpoint, requiring careful reading of descriptions.
All tools share the memory_ prefix, but the pattern after the prefix is inconsistent: verb_noun forms like memory_ingest_project and memory_init_project, bare verbs like memory_forget and memory_remember, and nouns like memory_engineering_context, memory_status, and memory_checkpoint. This is readable but not a single predictable convention.
Eight tools is well-scoped for a memory server covering project initialization, ingestion, remembering, checkpointing, recall, forgetting, and status. Each tool has a clear place in the workflow, and none feels redundant.
The memory lifecycle is largely covered: init, ingest, remember, checkpoint, recall, forget, and status. Minor gaps include no direct single-memory retrieval or search operation, and the relationship between memory_engineering_context and memory_recall is not fully clarified.