PseudoLife-MCP

Official

Overview Schema Source Related Servers Discussions Admin

memory_dream

Consolidate recent memories into canonical facts and knowledge graph structures. Manage the dream process with status, pull, run, deep, runs, and rollback actions.

Instructions

Drive the dream — consolidation of recent memories into canonical facts and graph structure.

Actions: status: backlog + whether a sweep would fire. Read-only. pull: unconsolidated memories, oldest first; write facts via memory_fact_set, then commit. run: a server-side dream with the configured extractor (loop to drain). deep: full-corpus graph consolidation; dry run unless apply. Settle candidates via memory_graph_review; duplicate lesson/world slots are listed for hand curation. runs: recent dream passes (tallies, status). rollback: revert a committed pass from its journal (facts + events; traces/cursor kept).

Returns: per-action dict; {error} on bad input.

Input Schema

TableJSON Schema

Name	Required	Description
`apply`	No	deep: True writes the consolidation (graph tables are snapshotted first); the default is a dry run.
`limit`	No	pull/run: how many memories to process (pull defaults to 40). runs: how many passes to list (defaults to 10).
`action`	Yes
`cursor`	No	commit: the newest pulled timestamp. Required for that action.
`run_id`	No	rollback: which pass to revert; defaults to the newest committed pass.
`snippets`	No	deep: False omits the evidence snippets.

Output Schema

TableJSON Schema

Name	Required	Description	Default
No arguments

Schema Changelog

Changes observed during successful MCP inspections.

Changed5 schema fields changedv0.15.0Sep 3, 2026

addedInput schema / properties / apply / description

Added value: +"deep: True writes the consolidation (graph tables are snapshotted first); the default is a dry run."

addedInput schema / properties / cursor / description

Added value: +"commit: the newest pulled timestamp. Required for that action."

addedInput schema / properties / limit / description

Added value: +"pull/run: how many memories to process (pull defaults to 40). runs: how many passes to list (defaults to 10)."

addedInput schema / properties / run_id / description

Added value: +"rollback: which pass to revert; defaults to the newest committed pass."

addedInput schema / properties / snippets / description

Added value: +"deep: False omits the evidence snippets."

Changed2 schema fields changedv0.14.0Aug 22, 2026

changedInput schema / properties / action / enum

Previous value: -[
-  "status",
-  "pull",
-  "commit",
-  "run",
-  "deep"
-]New value: +[
+  "status",
+  "pull",
+  "commit",
+  "run",
+  "deep",
+  "runs",
+  "rollback"
+]

addedInput schema / properties / run_id

Added value: +{
+  "anyOf": [
+    {
+      "type": "integer"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "title": "Run Id"
+}

First observedv0.11.0Jul 31, 2026

TDQS

A3.9/5.0

Behavior4/5

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

With no annotations, the description carries the behavioral burden and discloses side effects: status is read-only, pull writes facts via memory_fact_set and then commits, deep can write with a snapshot, and rollback reverts facts/events while keeping traces/cursor. It does not address idempotency or commit as a standalone behavior, but nothing contradicts annotations.

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

Conciseness5/5

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

The description front-loads a one-line purpose, then presents a scannable action list with no redundant prose. Every bullet adds operational information, and the return note is appropriately brief.

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

Completeness4/5

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

For a complex seven-action tool with no annotations, the description covers most behavioral and workflow context, and the output schema covers return values. It is slightly incomplete because the commit action has no bullet of its own and there is no direct comparison to memory_consolidate, but the schema fills the commit cursor requirement and the overall picture is actionable.

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 coverage is 83%, so the baseline is 3, but the description adds real value by decoding the action enum and linking parameter semantics to workflow (pull then commit, dry run unless apply, rollback targeting a pass). It leaves snippets to the schema, which already documents it, so the description complements rather than restates the schema.

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

Purpose4/5

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

The opening sentence states the tool's purpose — consolidating recent memories into canonical facts and graph structure — and the action list makes the operational scope concrete. It does not explicitly contrast itself with sibling tools like memory_consolidate, so it stops short of full sibling differentiation.

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

Usage Guidelines3/5

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

Each action bullet provides usable context, including status being read-only, deep being a dry run unless apply is set, and rollback reverting a committed pass. However, there is no when-not or selection guidance against likely alternatives such as memory_consolidate, so the agent must infer when this orchestration tool is preferred.

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