save-to-memory

Extracts and stores conversation knowledge into a Context Memory DB via MCP tools.

10|3|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/Cynthia1070711/PHYCOOL_Tools --skill save-to-memory-cynthia1070711
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: save-to-memory
Source: https://github.com/Cynthia1070711/PHYCOOL_Tools/tree/main/config-templates/claude/skills/save-to-memory
Command: npx skills add https://github.com/Cynthia1070711/PHYCOOL_Tools --skill save-to-memory-cynthia1070711

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Valuable decisions, analyses, bug fixes, and progress updates made during AI conversations are lost when the session ends. This Skill scans the current conversation, classifies reusable knowledge, and persists it into a long-term Context Memory DB so it can be recalled in future sessions. ## Core Features & Use Cases - Automatic Knowledge Extraction: Scans the conversation for decisions, architecture analyses, technical debt, bug fixes, success patterns, progress tracking, CR findings, workflow runs, and performance baselines, each with a priority level. - Typed MCP Tool Routing: Maps each knowledge type to the correct MCP tool (add_context, add_tech, add_cr_issue, log_workflow, upsert_benchmark) with the right category and parameter conventions. - Deduplication Before Writing: Queries search_context, search_tech, search_debt, search_stories, and related tools to skip or update existing records instead of creating duplicates. - Use Case: After a long debugging and design session, invoke the skill to automatically distill the root cause, the chosen architecture decision, and the current progress into structured memory records with tags and related files. ## Quick Start Ask the AI to save the current conversation to the memory database, optionally specifying a type such as decisions or progress.

Frequently Asked Questions about save-to-memory

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

FAQPage Schema
How do I save AI conversation content to a long-term memory database?

Invoke the save-to-memory skill, which scans the current conversation, classifies knowledge points by type, checks for duplicates, and batch-writes them into the Context Memory DB using MCP tools like add_context and add_tech.

What types of knowledge can be stored from a conversation?

The skill stores decisions, architecture analyses, analysis reports, technical debt, bug fixes, success patterns, progress tracking, code review findings, workflow execution records, and performance baselines, each routed to a matching MCP tool and category.

How does the skill prevent duplicate memory records?

Before writing, it queries type-specific search tools such as search_context, search_tech, search_debt, and search_stories. If a matching record exists, it skips the write or asks the user whether to update it.

Can I save only specific types of content like decisions or progress?

Yes. Passing a parameter such as decisions or progress restricts the scan to that knowledge type, while calling the skill without parameters scans the entire conversation for all supported types.

What content is not allowed to be written to the memory database?

Unverified speculative conclusions, passwords or keys, off-topic chatter, and raw unedited conversation dumps are forbidden. Content must be distilled, under 2000 characters per record, and carry at least two tags.