Skip to main content
Glama

Mnemoverse Memory

by io.github.mnemoverse
OverviewDiscussionsAdmin

memory_feedback

Report whether memories returned by memory_read were actually helpful. This is a learning signal, not a log: positive feedback raises a memory's ranking so it surfaces faster next time (across all of the user's tools), negative feedback lowers it so other memories out-rank it — nothing is erased and nothing decays with time. Call it right after you act on (or reject) recalled memories, passing the ids from the memory_read results as memory_ids. For memories read from a shared room, also pass that room's address as domain; your own memories need no domain. A read-only room member cannot rate the room's memories.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
domainNoOnly for memories read from a shared room: that room's address (xroom:...), exactly as you read it. Omit it for your own memories, which are rated by id alone.
outcomeYesHow helpful was this? 1.0 = very helpful, 0 = neutral, -1.0 = harmful/wrong
memory_idsYesIDs of the memories to rate, the `id:` line of each memory_read result.

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
avg_valenceNoAverage valence of the rated memories after the update. Reported as 0 in an asynchronous acknowledgement, where the real value is computed later — a 0 here is therefore not evidence of a neutral outcome.
updated_countYesHow many memories the service reports it applied the rating to. Processed synchronously this is the real count of memories that existed and were updated; processed asynchronously it is a best-effort ACCEPTED-count estimate — the number of IDs submitted — and the authoritative number is not known until the background worker runs. Zero means no submitted ID matched in the service's resolved request scope.
coactivation_edgesNoNumber of feedback-driven query/result concept co-activation edges changed by the service. This is separate from ordinary Hebbian strengthening among a memory's own concepts. This connector does not send query_concepts, so live calls through this tool report 0; asynchronous acknowledgements also report 0.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • removedInput schema / properties / atom_ids
      Removed value: -{
-  "description": "Deprecated since 0.11, removed in 0.13: atom_ids is the old name of memory_ids, still accepted on its own until then. Pass memory_ids instead.",
-  "items": {
-    "type": "string"
-  },
-  "minItems": 1,
-  "type": "array"
-}
    • changedInput schema / properties / memory_ids / description
      Previous value: -"Required: IDs of the memories to rate, the `id:` line of each memory_read result. (Optional in this schema only while the deprecated atom_ids is still accepted in its place.)"New value: +"IDs of the memories to rate, the `id:` line of each memory_read result."
    • changedInput schema / required
      Previous value: -[
-  "outcome"
-]New value: +[
+  "memory_ids",
+  "outcome"
+]
  2. Changed9 schema fields changed
    • removedInput schema / additionalProperties
      Removed value: -false
    • addedInput schema / properties / atom_ids
      Added value: +{
+  "description": "Deprecated since 0.11, removed in 0.13: atom_ids is the old name of memory_ids, still accepted on its own until then. Pass memory_ids instead.",
+  "items": {
+    "type": "string"
+  },
+  "minItems": 1,
+  "type": "array"
+}
    • addedInput schema / properties / domain
      Added value: +{
+  "description": "Only for memories read from a shared room: that room's address (xroom:...), exactly as you read it. Omit it for your own memories, which are rated by id alone.",
+  "type": "string"
+}
    • changedInput schema / properties / memory_ids / description
      Previous value: -"Personal-scope memory IDs from a prior memory_read. Shared-room IDs cannot currently be rated because this tool has no domain input."New value: +"Required: IDs of the memories to rate, the `id:` line of each memory_read result. (Optional in this schema only while the deprecated atom_ids is still accepted in its place.)"
    • removedInput schema / properties / memory_ids / items / format
      Removed value: -"uuid"
    • removedInput schema / properties / memory_ids / items / pattern
      Removed value: -"^([0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12})$"
    • removedInput schema / properties / memory_ids / maxItems
      Removed value: -20
    • changedInput schema / properties / outcome / description
      Previous value: -"Outcome signal: -1.0 (failure) to +1.0 (success)."New value: +"How helpful was this? 1.0 = very helpful, 0 = neutral, -1.0 = harmful/wrong"
    • changedInput schema / required
      Previous value: -[
-  "memory_ids",
-  "outcome"
-]New value: +[
+  "outcome"
+]
  3. First observed

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only say readOnlyHint=false, idempotentHint=false, destructiveHint=false; the description goes well beyond by explaining the ranking consequence of positive vs. negative feedback, that it applies across all of the user's tools, and that nothing is erased or time-decayed. It also discloses the authorization constraint that a read-only room member cannot rate room memories, which the annotations do not capture.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The core purpose and the ranking consequence are front-loaded in the first two sentences, with the mechanics (ids, domain, permission caveat) following. It is a little dense at four sentences, but every sentence carries distinct information and none is redundant with the schema.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With an output schema present, return values need no explanation, and the description still covers the trigger, the effects, the parameter sourcing, the domain rule, and the permission boundary. Nothing an agent needs to invoke this mutation correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the baseline is 3; the description adds real meaning on top by specifying where memory_ids come from (the id: line of memory_read results) and precisely when domain must be supplied versus omitted. It stops short of adding numeric guidance for the outcome range, which the schema already handles well.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The first sentence gives a specific verb and resource ('Report whether memories returned by memory_read were actually helpful') and immediately frames the tool's role as a learning signal rather than a log, which separates it from memory_read and memory_write in the sibling set. An agent can identify the tool's function without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives explicit timing ('Call it right after you act on (or reject) recalled memories') and an explicit prerequisite mapping ('passing the ids from the memory_read results as memory_ids'), plus the domain condition and the read-only-room exclusion. When-to-use, when-not, and the required upstream tool are all stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Mnemoverse Memory toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources