memory-schema

Enforces correct memory server tool call shapes, required fields, enums, and character caps before memory writes.

2|Updated Jul 24, 2026
One-click install
npx skills add https://github.com/konstantinos-malavazos/claude-code-playbook --skill memory-schema-konstantinos-malavazos
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: memory-schema
Source: https://github.com/konstantinos-malavazos/claude-code-playbook/tree/main/templates/skills/memory-schema
Command: npx skills add https://github.com/konstantinos-malavazos/claude-code-playbook --skill memory-schema-konstantinos-malavazos

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents writing to a memory MCP server frequently invent fields, nest arguments incorrectly, violate strict enums, or exceed character caps, causing repeated validation failures. This Skill auto-loads the exact call shape and rules before any memory write so writes succeed on the first attempt. ## Core Features & Use Cases - Call-shape enforcement: Documents the wrapper-trio dispatch pattern (execute_forgetful_tool with sibling tool_name and arguments) and the most common nesting failure. - Field, enum, and cap reference: Lists required fields per primitive (create_memory, create_document, create_entity, create_project, create_code_artifact), strict case-sensitive enums, and hard character caps with split-don't-trim guidance. - Asymmetric linking table and error playbook: Maps which link operations exist, which tools do not exist, and provides an error-to-fix lookup table for validation failures. - Use Case: Before calling create_memory, the agent loads this Skill, supplies all six required fields with importance as an integer, keeps content under 2000 characters, and sets project_ids so the memory is visible to project-scoped queries. ## Quick Start Load the memory schema before writing any memory and create a memory with title, content, context, keywords, tags, and integer importance using the documented call shape.

Frequently Asked Questions about memory-schema

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix 'unexpected keyword argument tool_name' in Forgetful MCP?

This error means tool_name was nested inside the arguments object. Pass tool_name and arguments as sibling parameters to execute_forgetful_tool, with the tool's fields inside arguments.

What fields are required for create_memory in Forgetful?

create_memory requires six fields: title (200 chars), content (2000 chars), context (500 chars), keywords (10 items), tags (10 items), and importance as an integer from 1 to 10. All six must be present or validation fails.

Why does create_entity fail with 'Input should be Organization'?

The entity_type field is a strict case-sensitive enum accepting only Organization, Individual, Team, Device, or Other. Values like Service, Database, or Application fail; use Other with a custom_type field for technical concepts.

How do I link an entity to a document in Forgetful?

Direct entity-to-document linking is not supported. Bridge it by naming the entity in the document content, sharing tags, and connecting both through a memory using link_entity_to_memory plus update_memory with document_ids.

What should I do when memory content exceeds 2000 characters?

Split the content into two or three atomic linked memories rather than trimming it. Each part should stand alone and reference the others in its context field, since trimming usually removes the reasoning.

Why is my memory never returned by project-scoped queries?

A memory created with an empty project_ids list is invisible to every project-scoped query. Always resolve project ids with list_projects at session start and set project_ids on every new memory.