UNITARES

Overview Schema Source Related Servers Discussions Admin

search_shared_memory

Search the cross-agent knowledge graph to find prior findings, excluding archived or cold rows unless explicitly included.

Instructions

Search the cross-agent knowledge graph for prior findings. Rows in status archived or cold are excluded unless you set status explicitly or pass include_archived / include_cold; a resolved or closed finding is still returned. Reading is not free of effect: every successful search appends a knowledge_read audit row naming the reader and a redacted copy of the query, which is why this tool is not annotated read-only. It serves unbound callers, so it works before start_session, unlike the writes: use store_finding to add a finding and use_tool(tool_name='update_finding', ...) to revise one.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`tags`	No	Exact any-of tag filter.
`limit`	No	Max results, 1-100 (more is capped).
`query`	No	Search text.
`status`	No	Filter by status: open, resolved, archived, superseded.
`sort_by`	No	search order: relevance, or created_at (newest matches first).
`agent_id`	No	Filter by author agent; agent_id_filter wins if both are set.
`operator`	No	Boolean operator for multi-term FTS queries
`semantic`	No	Legacy semantic on/off toggle.
`severity`	No	Filter by severity: low, medium, high, critical.
`search_mode`	No	auto, or force fts/semantic/hybrid.
`include_cold`	No	Include cold-storage rows.
`created_after`	No	search: created after ISO time.
`response_mode`	No	Read-envelope mode. Default lean: one-line digests; compact adds diagnostics, full adds raw_governance.	lean
`authority_mode`	No	prefer_governed (default) down-ranks imported memory; all keeps raw order.
`created_before`	No	search: created before this ISO time.
`discovery_type`	No	Filter by discovery type, e.g. bug_found.
`min_similarity`	No	Semantic similarity floor.
`agent_id_filter`	No	Author agent UUID filter; wins over agent_id.
`include_details`	No	Inline details need response_mode='full'; else open one with knowledge(action='details').
`continuity_token`	No	Same-process rebind proof only; never a cross-process resume.
`include_archived`	No	Include archived rows.
`client_session_id`	No	Binding id for calls in this process; not a cross-process proof.
`include_provenance`	No	Add provenance fields.
`exclude_agent_labels`	No	Omit search results whose writer display label matches one of these values
`recency_half_life_days`	No	search: score halves per N days.

Schema Changelog

Changes observed during successful MCP inspections.

Changed3 schema fields changedv3.2.0Oct 5, 2026

changedInput schema / properties / agent_id / description

Previous value: -"Filter by agent (for action=get, search; omit when using discovery_id readback)"New value: +"Filter by author agent; agent_id_filter wins if both are set."

changedInput schema / properties / query / description

Previous value: -"Search query (for action=search)"New value: +"Search text."

changedInput schema / properties / status / description

Previous value: -"Status filter/update value (open, resolved, archived, superseded)"New value: +"Filter by status: open, resolved, archived, superseded."

Changed37 schema fields changedv3.1.0Sep 30, 2026

changedInput schema / description

Previous value: -"Parameters for knowledge"New value: +"Knowledge graph parameters."

addedInput schema / properties / agent_id_filter

Added value: +{
+  "anyOf": [
+    {
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Author agent UUID filter; wins over agent_id."
+}

addedInput schema / properties / authority_mode

Added value: +{
+  "anyOf": [
+    {
+      "enum": [
+        "prefer_governed",
+        "all"
+      ],
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "prefer_governed (default) down-ranks imported memory; all keeps raw order."
+}

changedInput schema / properties / client_session_id / description

Previous value: -"In-session binding id from start_session()/identity(); pass it on same-process calls. Not a cross-process proof."New value: +"Binding id for calls in this process; not a cross-process proof."

removedInput schema / properties / closure_class

Removed value: -{
-  "anyOf": [
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Closing standard for action=update: fix_verified | unobserved | not_reproducible | obsolete | duplicate. 'fix_verified' needs a deployed change whose effect was observed."
-}

removedInput schema / properties / closure_evidence

Removed value: -{
-  "anyOf": [
-    {
-      "additionalProperties": true,
-      "type": "object"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Evidence for closure_class. Required keys: fix_verified needs {deployed, observed}; unobserved needs {window, instrument_check}."
-}

removedInput schema / properties / confidence

Removed value: -{
-  "anyOf": [
-    {
-      "type": "number"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Writer-supplied confidence for action=store, validated to 0-1; not independently verified or adjusted by governance metrics"
-}

changedInput schema / properties / continuity_token / description

Previous value: -"Ownership proof from onboard()/identity(), for same-live-process rebinds only. Not a cross-process resume credential."New value: +"Same-process rebind proof only; never a cross-process resume."

addedInput schema / properties / created_after

Added value: +{
+  "anyOf": [
+    {
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "search: created after ISO time."
+}

addedInput schema / properties / created_before

Added value: +{
+  "anyOf": [
+    {
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "search: created before this ISO time."
+}

changedInput schema / properties / discovery_type / description

Previous value: -"action=store; one of architectural_decision, learning, pattern, bug_fix, refactoring, documentation, experiment, question, note, rule, insight, bug_found, bug, improvement, exploration, observation."New value: +"Filter by discovery type, e.g. bug_found."

removedInput schema / properties / dry_run

Removed value: -{
-  "anyOf": [
-    {
-      "type": "boolean"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Dry run mode (for action=cleanup, synthesize)"
-}

removedInput schema / properties / epoch_scope

Removed value: -{
-  "anyOf": [
-    {
-      "enum": [
-        "current",
-        "all"
-      ],
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Stats/list scope: current epoch only or all epochs"
-}

changedInput schema / properties / include_archived / description

Previous value: -"Include archived discoveries in search results (default: excluded)"New value: +"Include archived rows."

changedInput schema / properties / include_cold / description

Previous value: -"Include cold-storage (long-term) discoveries in search results (default: excluded)"New value: +"Include cold-storage rows."

changedInput schema / properties / include_details / description

Previous value: -"Expand results inline only with response_mode='full'; otherwise open one with knowledge(action='details')."New value: +"Inline details need response_mode='full'; else open one with knowledge(action='details')."

changedInput schema / properties / include_provenance / description

Previous value: -"Include provenance and lineage chain fields in search/details results"New value: +"Add provenance fields."

removedInput schema / properties / include_response_chain

Removed value: -{
-  "anyOf": [
-    {
-      "type": "boolean"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Include typed response chain for action=details"
-}

removedInput schema / properties / including_cold

Removed value: -{
-  "anyOf": [
-    {
-      "type": "boolean"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Include cold-storage discoveries in action=list raw status aggregates"
-}

removedInput schema / properties / length

Removed value: -{
-  "anyOf": [
-    {
-      "type": "integer"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Maximum details characters returned for action=details"
-}

changedInput schema / properties / limit / description

Previous value: -"Max results (for action=search: min 1, values above 100 are capped, 0 or negative is rejected)"New value: +"Max results, 1-100 (more is capped)."

removedInput schema / properties / max_chain_depth

Removed value: -{
-  "anyOf": [
-    {
-      "type": "integer"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Maximum response-chain traversal depth for action=details"
-}

removedInput schema / properties / memory_context

Removed value: -{
-  "anyOf": [
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "S22 provenance: memory/KG/transcript surfaces visible to the writer"
-}

removedInput schema / properties / min_members

Removed value: -{
-  "anyOf": [
-    {
-      "type": "integer"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Minimum discoveries a topic needs before it is rolled up (for action=synthesize, default 3)"
-}

changedInput schema / properties / min_similarity / description

Previous value: -"Minimum cosine similarity for semantic retrieval modes"New value: +"Semantic similarity floor."

removedInput schema / properties / offset

Removed value: -{
-  "anyOf": [
-    {
-      "type": "integer"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Character offset for action=details pagination"
-}

addedInput schema / properties / recency_half_life_days

Added value: +{
+  "anyOf": [
+    {
+      "exclusiveMinimum": 0,
+      "maximum": 36500,
+      "type": "number"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "search: score halves per N days."
+}

removedInput schema / properties / scope

Removed value: -{
-  "anyOf": [
-    {
-      "enum": [
-        "open",
-        "all",
-        "by_agent"
-      ],
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "KG audit scope for action=audit"
-}

changedInput schema / properties / search_mode / description

Previous value: -"Force retrieval mode for action=search. 'semantic' and 'hybrid' fail honestly when unsupported by the active backend."New value: +"auto, or force fts/semantic/hybrid."

changedInput schema / properties / semantic / description

Previous value: -"Legacy action=search toggle to force or skip semantic retrieval when supported"New value: +"Legacy semantic on/off toggle."

changedInput schema / properties / severity / description

Previous value: -"Severity: low, medium, high, critical (for action=store or action=update)"New value: +"Filter by severity: low, medium, high, critical."

addedInput schema / properties / sort_by

Added value: +{
+  "anyOf": [
+    {
+      "enum": [
+        "relevance",
+        "created_at"
+      ],
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "search order: relevance, or created_at (newest matches first)."
+}

changedInput schema / properties / tags / description

Previous value: -"Tags (action=store, note); for action=search an exact any-of filter in every search mode."New value: +"Exact any-of tag filter."

removedInput schema / properties / top_n

Removed value: -{
-  "anyOf": [
-    {
-      "type": "integer"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Maximum stale entries returned by action=audit"
-}

removedInput schema / properties / topic

Removed value: -{
-  "anyOf": [
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Synthesize just this one tag/topic (for action=synthesize). Omit to sweep the densest topics."
-}

removedInput schema / properties / use_llm

Removed value: -{
-  "anyOf": [
-    {
-      "type": "boolean"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Use the local LLM for the rollup narrative (for action=synthesize, default true; falls back to deterministic when unreachable)"
-}

removedInput schema / properties / use_model

Removed value: -{
-  "anyOf": [
-    {
-      "type": "boolean"
-    },
-    {
-      "type": "string"
-    },
-    {
-      "type": "null"
-    }
-  ],
-  "default": null,
-  "description": "Use the local model to assess stale entries for action=audit"
-}

First observedSep 8, 2026

TDQS

A4.7/5.0

Behavior5/5

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

Annotations declare readOnlyHint=false without explaining why; the description supplies exactly that missing context, disclosing that every successful search appends a knowledge_read audit row naming the reader plus a redacted query copy. It also documents the visibility side effect that archived/cold rows are hidden unless status/include_archived/include_cold are set, which is behavioral information the annotations cannot convey.

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?

Four tightly packed sentences, front-loaded with purpose then caveats, side effects, and routing, with no filler. It is dense rather than bloated, though the 'resolved or closed' clause sits slightly awkwardly against the status enum, which costs a point.

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 25-parameter, no-required, no-output-schema search tool, the description covers the high-risk behaviors: default filtering, the audit side effect, call ordering relative to start_session, and sibling routing. Response envelope modes and pagination are left to the schema, which documents them well, so remaining gaps are minor.

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 100%, so the baseline is 3, but the description adds real cross-parameter semantics: the interaction between the default status filter and include_archived/include_cold, and that resolved/closed findings still surface. It does not cover the 25 parameters exhaustively, but it lifts the most error-prone filtering logic above what the schema alone states.

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?

Opens with a specific verb and resource ('Search the cross-agent knowledge graph for prior findings') and explicitly separates itself from the write siblings by naming store_finding and update_finding. An agent can distinguish read-from-search versus add/revise without opening any 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?

States when it applies ('serves unbound callers, so it works before start_session, unlike the writes') and names the alternatives for adjacent operations (store_finding to add, use_tool(update_finding) to revise). The archived/cold exclusion rule and the resolved/closed exception give concrete filtering conditions rather than leaving them to inference.

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

Deploy Server

search_shared_memory

Instructions

Input Schema

Schema Changelog

TDQS

Other Tools