Skip to main content
Glama
LamantinAI

kaeru memory

by LamantinAI
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Rust
Hybrid

link

Create a directed, typed, weighted edge between two memory nodes by name or ID to define causes, contradictions, or supersessions; weight controls how knowledge chains route.

Instructions

Create a typed edge between two nodes (by name or id). Endpoints resolve in the active initiative first, then across all initiatives, so a link may span initiatives. Edge type defaults to refers_to. weight (0..1) is REQUIRED — it is HOW LOAD-BEARING the edge is and the only signal knowledge chains route on (path cost is 1−weight); there is no default because an unweighted graph makes every chain rank on noise. State it by the scale: 0.9–1.0 load-bearing (a cause, a source a conclusion rests on, a supersession — the edges a chain should follow); 0.6–0.8 supporting but not decisive; 0.3–0.5 loose / associative. DIRECTION matters for supersedes: link <new> <old> --edge_type supersedes — the replacement points at what it replaced, so an INBOUND supersedes means this node is obsolete.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
toYesDestination node name or id — resolved in the active initiative first, then across all initiatives.
fromYesSource node name or id — resolved in the active initiative first, then across all initiatives, so an edge may span initiatives.
cloudNoWhich cloud to mirror this edge change to, when BOTH endpoints are already shared. Omit with one cloud configured (it is unambiguous) or when the endpoints are local. With several configured and none named, the local edit still happens and the result says the edge was not mirrored.
weightYesConnection strength in `0..1` — REQUIRED, no default. How load-bearing this edge is, and the only signal knowledge chains route on (path cost is `1 − weight`). State it deliberately by how much the connection matters: 0.9–1.0 load-bearing (a cause, a source a conclusion rests on, a supersession — the edges a chain should follow); 0.6–0.8 supporting, not decisive; 0.3–0.5 loose / associative. There is no neutral fallback on purpose: an unweighted graph is what made every chain rank on noise.
edge_typeNoEdge type — a CLOSED vocabulary, one of exactly these: `refers_to` (default), `causal`, `derived_from`, `contradicts`, `part_of`, `blocks`, `targets`, `supersedes`, `verifies`, `falsifies`, `temporal`, `consolidated_to`. `supersedes` is directed: `src` supersedes `dst` — the replacement is the `src`, so an inbound `supersedes` means the node is obsolete. Nothing else is accepted (`related_to` and friends are not edge types). Snake_case or kebab-case both accepted.refers_to
initiativeNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.7.5

TDQS

A4.3/5.0
Behavior4/5

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

With no annotations, the description carries the full burden and does much of it: it discloses endpoint resolution order (active initiative first, then all), the possibility of cross-initiative edges, that weight has no default and why, and the directed semantics of supersedes. It omits permission/auth requirements, failure behavior when a name fails to resolve, and whether an existing edge is overwritten or duplicated.

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?

Front-loaded with purpose, then layered with resolution rules, weight guidance, and direction — a sensible order. It is dense but mostly earns its length; the duplication of the weight-scale explanation (already in the schema) is the main wasted sentence.

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 6-param mutation tool with no annotations and no output schema, the description covers the decisive behaviors: resolution order, weight semantics, edge vocabulary, direction. Remaining gaps are the undocumented `initiative` param, failure/resolution-error behavior, and what the call returns.

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 already 83%, so the baseline is 3; the description goes beyond it with the actionable weight bands and the explicit supersedes direction rule. That said, much of the weight and edge_type guidance is near-verbatim repetition of the schema text, and the `initiative` parameter is left undocumented in both places.

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?

States a precise verb+resource ('Create a typed edge between two nodes') and immediately pins down scope with the name-or-id resolution rule and cross-initiative possibility. An agent can distinguish this from siblings like unlink, reweight, or link_cloud without opening the schema, since the description itself contrasts edge creation against edge mutation and cloud mirroring.

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

Usage Guidelines4/5

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

Gives strong in-tool guidance — the weight band rubric, the required-vs-default rationale, and the explicit direction convention for supersedes (

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

Deploy Server

Other Tools