skills.rest
clacky-ai
clacky-aiOFFICIAL
Explore all skills by this author →
WebsiteGitHub

persist-memory

Persists conversation facts to Markdown memory files with topic merging and frontmatter.

1.2k|105|Updated Dec 30, 2025
Search Tags:#persistence#Markdown#yaml frontmatter#Memory#long-term storage#Agent#topic merging
One-click install
npx skills add https://github.com/clacky-ai/openclacky --skill persist-memory
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: persist-memory
Source: https://github.com/clacky-ai/openclacky/tree/main/lib/clacky/default_skills/persist-memory
Command: npx skills add https://github.com/clacky-ai/openclacky --skill persist-memory

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Important facts from conversations are easily lost once a session ends. This Skill writes durable, topic-organized memory files to ~/.clacky/memories/ so key information survives across sessions without manual note-taking.

Core Features & Use Cases

  • Topic-based file organization: Matches new information to existing memory files by topic, or creates a new slug-named Markdown file when no topic fits.
  • Merging and trimming: Integrates new facts into existing files, drops stale content, and enforces a soft size target of around 4000 characters per file.
  • YAML frontmatter management: Every memory file carries a topic and description in frontmatter for consistent discovery.
  • Use Case: After a long debugging session, the user asks the agent to remember the deployment target and code style preferences; the Skill updates deployment-target.md and code-style-preferences.md accordingly.

Quick Start

Ask the agent to remember the project's deployment target and preferred code style for future sessions.

Frequently Asked Questions about persist-memory

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

FAQPage Schema
How do I make an AI agent remember information across sessions?▼

Ask the agent to remember or note the information, and this Skill writes it to a topic-organized Markdown file under ~/.clacky/memories/. Each file uses YAML frontmatter with a topic and description so it can be found and updated later.

How are memory files organized and named?▼

Each memory file covers one topic and is named with a lowercase hyphen-separated slug, such as deployment-target.md. New information is merged into an existing matching topic file whenever possible rather than creating duplicates.

What format do the persisted memory files use?▼

Memory files are plain Markdown with YAML frontmatter containing a topic and one-line description, followed by concise factual Markdown content. The target content size is around 4000 characters per file.

What happens when a memory file gets too large?▼

When a file grows well past 8000 characters, the least important information is trimmed rather than splitting the topic across multiple files. Moderate overshoot of the 4000-character target is accepted without iterative rewriting.

Can this Skill decide on its own what to remember?▼

No. It is a pure executor subagent: the caller decides what must be written, and the Skill only handles correct file selection, merging, and writing. It is not user-invocable and runs as a forked agent.

Related Skills

Explore More Software Engineering Skills→
derrybirkettderrybirkett
1

note-that

Append categorized timestamped observations to notes/insights.md during conversations.

Community
Basic
iBz-04iBz-04
63

continual-learning

Capture agent experiences and persist learnings across sessions via databases and memory files.

Community
Advanced
ajayjohnajayjohn
1

learn

Extract durable insights from conversations and learning content into persistent memory.

Community
Advanced
Mark393295827Mark393295827
133

knowledge-ops

Deduplicate, classify, and audit multi-layer knowledge bases for retrieval readiness.

Community
Advanced