Ebb
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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 | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| rememberA | Store a piece of knowledge (note, decision, fact, person, client, ...). Returns its node id. Re-calling with the same node_id updates it. Set pinned=True for durable knowledge that should never be auto-archived. |
| connectA | Create a connection from node |
| reinforceC | Strengthen and refresh an existing connection you just used again — the spaced-repetition signal that keeps knowledge from decaying. |
| recallA | Retrieve the k most relevant nodes for a query. Ranking blends text match with recency-weighted activation. Archived (cold/frozen) nodes are excluded unless include_cold=True. Each result explains why it surfaced. |
| neighborsC | List the direct connections of a node. |
| forgetA | Permanently delete a node and its connections. Prefer letting decay + the tier-4 review queue handle phase-out; use this only for deliberate deletion. |
| pinB | Pin (or unpin) a node so it is never auto-archived — a human 'principal' mark on durable knowledge. |
| maintainA | Run the decay/archival pass: recompute activation, reassign tiers 1-4, and return the changes plus the tier-4 review queue (items awaiting human sign-off before deletion). |
| review_queueC | The tier-4 queue: knowledge that has decayed to the point of proposed deletion and is waiting for a human to confirm or reprieve it. |
| statsB | Graph size, tier distribution, and the most-active nodes right now. |
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 10 tools
Each tool serves a distinct operation: node creation/update, edge creation, edge reinforcement, query, neighbor listing, deletion, pinning, maintenance pass, review queue access, and statistics. No two tools overlap in purpose.
All tool names are single lowercase words, with most being verbs (remember, connect, reinforce, recall, forget, pin, maintain) and a few nouns (neighbors, review_queue, stats). The style is consistent but mixes verb and noun forms, a minor deviation from a pure verb-noun pattern.
With 10 tools, the server is well-scoped for a knowledge graph with memory decay and archival. Each tool earns its place, covering creation, linking, querying, deletion, and maintenance without unnecessary bloat.
The surface covers the full node lifecycle (remember/forget), edge operations (connect/reinforce), querying (recall/neighbors), and archival (pin/maintain/review_queue). Minor gaps exist—no explicit edge deletion and no direct node fetch by ID—but recall and forget adequately work around these.