Rank private memory candidates; defaults to active/abstract. Use detail=full for complete bodies within max_tokens. preview_truncated marks clipped previews; ranking and review_due are not truth judgments. Private to this agent identity.
| Name | Required | Description | Default |
|---|---|---|---|
| data | Yes |
| Name | Required | Description | Default |
|---|---|---|---|
| error | No | ||
| replayed | No | ||
| operation | No | ||
| truncated | No | ||
| next_cursor | No |
Changes observed during successful MCP inspections.
Input schema / properties / data / properties / detailAdded value: +{
+ "default": "abstract",
+ "description": "full returns complete note bodies within max_tokens; abstract includes preview_truncated. A full body is never silently shortened.",
+ "enum": [
+ "abstract",
+ "full",
+ "metadata"
+ ],
+ "type": "string"
+}Output schema / (root)Previous value: -nullNew value: +{
+ "additionalProperties": true,
+ "description": "Structured result for recall_memory. Resource fields vary by detail, permissions and operation. Paginated results use items/next_cursor; operation-key replays use operation/replayed. Failures have error and the MCP result has isError=true. See get_help for this tool's contract.",
+ "properties": {
+ "error": {
+ "properties": {
+ "code": {
+ "type": "string"
+ },
+ "help": {
+ "type": "string"
+ },
+ "message": {
+ "type": "string"
+ },
+ "next_action": {
+ "type": "string"
+ },
+ "retry_after_seconds": {
+ "type": "integer"
+ }
+ },
+ "required": [
+ "code",
+ "message",
+ "next_action",
+ "help"
+ ],
+ "type": "object"
+ },
+ "next_cursor": {
+ "type": "string"
+ },
+ "operation": {
+ "type": "object"
+ },
+ "replayed": {
+ "type": "boolean"
+ },
+ "truncated": {
+ "type": "boolean"
+ }
+ },
+ "type": "object"
+}Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover the safety profile (readOnly, non-destructive, closed-world), yet the description adds real behavioral context beyond them: that results are private to this agent identity, that preview_truncated marks clipped previews, and that ranking, review_due, and previews are not truth judgments. Those are semantic caveats an agent could not derive from the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loads the core action and default in the first clause, then adds qualifying caveats in compact semicolon-separated statements. Dense and efficient, though 'review_due' is referenced without appearing in the input schema, which makes one clause slightly opaque.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
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 not be described, and the description still flags the notable response markers (preview_truncated, ranking not being truth) plus the privacy boundary. It is nearly complete for a read-only recall tool, missing only explicit guidance on choosing it over sibling retrieval tools.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Reported schema coverage is 0% at the top level, though the nested properties carry their own enum/default descriptions. The description adds value for detail=full and its interaction with max_tokens, but says nothing about status, tags, limit, or query semantics, so it only partially compensates for the wrapper's missing documentation.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource ('Rank private memory candidates') plus the default scope (active/abstract), so the agent knows this ranks memory rather than fetching one entry. It does not name or differentiate itself from nearby siblings like get_memory, list_memory, or search, so the distinction rests on inference.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives operational hints ('Use detail=full for complete bodies within max_tokens') and states the defaults, which implies when the more expensive mode is warranted. But there is no explicit when-to-use versus get_memory/list_memory/search, and no stated exclusions, leaving tool selection to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.