Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
COGVAULT_LOGNoSet to 'off' to disable effectiveness logging, or provide a file path to relocate the log. Defaults to enabled logging in the tenant directory.
COGVAULT_MODELNoEmbedding model to use. Defaults to paraphrase-multilingual-MiniLM-L12-v2.paraphrase-multilingual-MiniLM-L12-v2
COGVAULT_TENANTNoPath to the memory directory for this agent (tenant). Defaults to ~/.cogvault/memory.~/.cogvault/memory

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
cogvault_recallA

Search this agent's persistent memory. Pass a natural-language query; returns the most relevant memory snippets (hybrid semantic + keyword).

cogvault_recordA

Save a fact to persistent memory as a Markdown card. It becomes searchable on the next recall. Pass type so the card can be filtered on recall, and title so it gets a meaningful filename.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A3.7/5.0

Scored across 2 tools

Disambiguation5/5

The two tools are a clean read/write pair: recall searches memory while record writes to it. There is no plausible way to confuse them, and their descriptions reinforce the distinction.

Naming Consistency5/5

Both tools use the identical `cogvault_` prefix followed by a single clear verb (recall, record). The pattern is perfectly predictable and readable.

Tool Count3/5

Two tools is a thin surface even for a focused memory store; the domain of persistent memory naturally invites a few more operations. It is defensible as a minimal core, but sits below the well-scoped 3-15 range.

Completeness3/5

The server covers create and read but has no update, delete/forget, or list operations, leaving common memory-lifecycle tasks (correcting a stale fact, removing sensitive data) as dead ends. Core recall/record workflows do work.

Maintenance

ActivityNo data
ResponsivenessNo issues