Conversation
There was a problem hiding this comment.
Pull request overview
Adds a new “nano memory” skill intended to guide AI agents in maintaining persistent, file-based memory across stateless sessions.
Changes:
- Introduces a new
skills/nano-memory/SKILL.mddefining a file-based memory architecture (daily logs + curated long-term memory). - Documents memory operations (search/read/write/edit) and SOP-style workflows for retrieval and maintenance.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| ### Core Memory Files | ||
| - **`memory/YYYY-MM-DD.md` (Daily Logs):** Raw, chronological logs of what happened. Use this for daily tasks, scratchpad thinking, and immediate context. | ||
| - **`MEMORY.md` (Long-Term Memory):** Your curated, distilled knowledge base. Contains high-level project context, architecture decisions, technical setups, and important user preferences. | ||
|
|
There was a problem hiding this comment.
The skill defines memory/YYYY-MM-DD.md and MEMORY.md as core files but never specifies where they should live (project root vs skill directory vs home directory) or how to initialize them. This ambiguity can lead to writing memory into an unintended working directory. Please document the expected base directory and add a brief initialization step (e.g., ensure memory/ exists and create MEMORY.md if missing).
| ### Core Memory Files | |
| - **`memory/YYYY-MM-DD.md` (Daily Logs):** Raw, chronological logs of what happened. Use this for daily tasks, scratchpad thinking, and immediate context. | |
| - **`MEMORY.md` (Long-Term Memory):** Your curated, distilled knowledge base. Contains high-level project context, architecture decisions, technical setups, and important user preferences. | |
| ### Core Memory Files | |
| All memory files live in the project root directory (the same directory that contains this skill file), unless explicitly documented otherwise. | |
| - **`memory/YYYY-MM-DD.md` (Daily Logs):** Raw, chronological logs of what happened. Use this for daily tasks, scratchpad thinking, and immediate context. | |
| - **`MEMORY.md` (Long-Term Memory):** Your curated, distilled knowledge base. Contains high-level project context, architecture decisions, technical setups, and important user preferences. | |
| **Initialization (run once per project):** | |
| - Ensure the `memory/` directory exists at the project root. If it does not exist, create it before writing any daily logs. | |
| - Ensure `MEMORY.md` exists at the project root. If it does not exist, create an empty `MEMORY.md` (or a simple heading like `# Long-Term Memory`) before attempting to read or edit it. |
| You have access to native file tools and standard OS commands. Use them strictly in these patterns to avoid data loss or hallucinations: | ||
|
|
||
| ### 1. 🔍 Search (OS-Specific Commands) | ||
| - **When to use:** ALWAYS use this before answering questions about past work, decisions, or dates to find where information is stored. | ||
| - **Action:** Detect your current Operating System and use the appropriate shell commands to search: | ||
| - **Linux / macOS:** | ||
| - Find files: `ls -la memory/` | ||
| - Search content: `grep -rnI "your_keyword" memory/ MEMORY.md` |
There was a problem hiding this comment.
This doc hard-codes tool names (read_file, write_file, edit_file) and assumes shell command execution for ls/grep/dir/findstr, but the repository README advertises skills for multiple assistants (Claude Code, Cursor, etc.) where tool names/capabilities differ. To keep the skill portable, describe these as abstract capabilities (read/write/edit + optional shell/terminal search) and note that the agent should map them to the host environment’s equivalent tools.
| @@ -0,0 +1,71 @@ | |||
| --- | |||
| name: "nano memory skill" | |||
There was a problem hiding this comment.
The front matter metadata is inconsistent with the existing skill’s format and may be harder to consume reliably: the other skill uses a slug-style name (no spaces/quotes) and includes a version field. Consider aligning name to the directory/skill id (e.g., nano-memory) and adding a version key for traceability/versioning consistency.
| name: "nano memory skill" | |
| name: nano-memory | |
| version: 1.0.0 |
Summary
This PR introduces a new Nano Memory Skill that provides comprehensive guidelines and workflows for AI agents to maintain persistent memory using native file operations and standard OS commands. This skill enables stateless sessions to retain context, technical decisions, and historical information through structured file-based memory management.
Changes Made
This PR enhances the AgentScope skills ecosystem by providing a robust foundation for persistent memory management, enabling more intelligent and context-aware AI assistant interactions.