Agent Memory Skill

A lightweight persistent memory workflow for this repository.

Location

• •Memory root: .github/skills/agent-memory/memories/

When to use

Use this skill when the user asks to:

• •remember / save / note something
• •recall past decisions or findings
• •organize or clean previous notes

Also use proactively when you find knowledge worth preserving:

• •non-obvious decisions and rationale
• •tricky fixes and gotchas
• •research findings that took effort
• •resumable in-progress context

Privacy rule for this project

• •Treat bio/ as sensitive.
• •Save only the minimum necessary summary.
• •Do not copy raw personal details unless explicitly requested.

File format

Store each memory as Markdown with frontmatter:

---
summary: "1-2 lines: what this memory contains and why it matters"
created: 2026-02-16
updated: 2026-02-16
status: in-progress # in-progress | pending | archived
tags: [strategy]
related: [source/trend/trend.md]
---

summary and created are required.

Suggested structure

memories/
├── project-context/
├── trend/
├── strategy/
└── content-ops/

Use kebab-case for folder/file names.

Search workflow

# list summaries quickly
rg "^summary:" .github/skills/agent-memory/memories/ --no-ignore --hidden

# filter by keyword
rg "^summary:.*keyword" .github/skills/agent-memory/memories/ --no-ignore --hidden -i

# filter by tags
rg "^tags:.*keyword" .github/skills/agent-memory/memories/ --no-ignore --hidden -i

Quality guidelines

• •Write for future resumption (context, state, next step).
• •Keep summaries decisive.
• •Update stale notes or remove obsolete ones.