Guide · Agent skill
How to use the Memory Status Dashboard agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The memory-status skill is a read-only health dashboard for the persistent memory an AI coding agent keeps for a project. It counts lines in the agent's memory and rules files, shows how close they are to size limits, flags references to files that no longer exist, and suggests cleanup steps. This guide explains what an agent skill is, how to install this one, how to trigger it, and what to watch for.
Memory Status Dashboardalirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Confirm you use persistent agent memory. The skill needs a Claude memory directory under ~/.claude/projects. Without one, there is nothing to report.
- 2Download and unzip the skill. Download the ZIP from this page and unzip it so SKILL.md sits in a memory-status folder.
- 3Place the folder. Put it in ~/.claude/skills/ for all projects, or in the project's .claude/skills/ for one project. You can also upload it under Settings → Capabilities → Skills.
- 4Run the dashboard. In your project, ask for /si:memory-status or ask how full the agent memory is.
- 5Read the capacity rating. Compare line counts to the thresholds. Green is under 60% full, yellow is 60-90%, red is above 90%.
- 6Act on recommendations. For yellow or red capacity, run /si:memory-review. Delete stale references to missing files.
What an agent skill is and what memory-status does
An agent skill is a folder containing a SKILL.md file that gives an AI agent extra instructions for a specific task. When a request matches the skill's description, the agent follows its steps.
This skill inspects the agent's memory for a project. It locates the auto-memory directory under ~/.claude/projects, counts lines in MEMORY.md and CLAUDE.md, lists topic files, counts rules files in .claude/rules/, and checks MEMORY.md for file paths that no longer exist. It then rates capacity as healthy, warning or critical and gives recommendations.
- Line counts for MEMORY.md, CLAUDE.md and ~/.claude/CLAUDE.md
- Topic file list and number of rules files
- Capacity rating against the thresholds in the skill
- Stale references (files that no longer exist) and duplicate entries
- A --brief one-line summary mode
Install the memory-status skill
Download the ZIP for this skill from this page and unzip it. The skill is a folder that contains SKILL.md.
For a personal install that works in all your projects, unzip the folder into ~/.claude/skills/. To limit it to one project, unzip it into that project's .claude/skills/ folder instead.
You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Single project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills, then upload
Trigger the skill with commands or plain requests
The skill responds to the command /si:memory-status. The source lists two forms: the full dashboard and a one-line brief summary.
The skill also activates when you ask how full or healthy the agent memory is. You do not need the exact command for that kind of question.
- Full dashboard: /si:memory-status
- One-line summary: /si:memory-status --brief
- Natural request: "How full is the agent memory for this project?"
/si:memory-status
/si:memory-status --briefRead the dashboard and capacity thresholds
The skill rates MEMORY.md and other files against fixed line thresholds. Use the table to interpret the result.
For MEMORY.md, the dashboard shows the count against a 200-line limit. Overall capacity is shown as green below 60% full, yellow from 60% to 90%, and red above 90%. In red, the source warns that auto-memory may start dropping older entries.
- MEMORY.md: healthy under 120 lines, warning 120-180, critical above 180
- CLAUDE.md: healthy under 150 lines, warning 150-200, critical above 200
- Topic files: healthy 0-3, warning 4-6, critical above 6
- Stale entries: healthy 0, warning 1-3, critical above 3
Clean up what the report finds
The skill only reports and recommends. It does not change your memory files. When capacity is yellow or red, run /si:memory-review, which the source names as the separate command for promoting or cleaning up memory.
Stale references are file paths in MEMORY.md that no longer point to a real file. Delete those lines, then run the status check again to confirm the count drops.
- Yellow: run /si:memory-review to identify promotion candidates
- Red: run /si:memory-review now
- Stale refs: delete lines that point to missing files
/si:memory-reviewSafety notes: scripts and network
This skill runs scripts. It uses a bash shell with standard tools such as wc, ls, grep and sed. It does not use the network, according to the skill's requirements.
Because it runs shell commands, review SKILL.md before installing if you want to see exactly what will run. The commands only read files and print results.
- Runs scripts: yes (bash with wc, ls, grep, sed)
- Uses network: no
- Read-only: it reports and recommends but does not edit memory
What people use it for
- Checking whether MEMORY.md is nearing the 200-line limit before a long project session
- Finding file paths in memory that point to deleted or renamed files
- Keeping CLAUDE.md under its size thresholds in a team repository
- Getting a quick one-line status with /si:memory-status --brief before starting work
Frequently asked questions
How do I install the memory-status skill?
Download the ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ for all projects, or in a project's .claude/skills/ folder for one project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
What do I need to run memory-status?
You need a bash shell with standard tools (wc, ls, grep, sed) and Claude Code or a compatible agent that has a Claude memory directory under ~/.claude/projects. The skill runs scripts and does not use the network.
Is memory-status free?
The skill is released under the MIT licence, according to this listing. Check the licence file in the repository for the full terms.
Does memory-status change or delete my memory files?
No. The skill only reports line counts, capacity, stale references and recommendations. The cleanup is done with the separate /si:memory-review command.
Why does the dashboard say my memory is in the red?
Red means above 90% of capacity, or MEMORY.md above 180 lines in the thresholds. The source warns that auto-memory may start dropping older entries, so run /si:memory-review and remove stale or duplicate content.
What can I use instead of memory-status?
If you do not use persistent agent memory, this skill has nothing to report on. You can check CLAUDE.md and rules files manually with wc -l, which is the same line count the skill uses.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you work with an AI coding agent that keeps persistent project memory and you want a quick read on whether those notes are getting too large or pointing at missing files. Skip it if you do not use persistent agent memory, or if you need the skill itself to edit memory, since it only reports and recommends.
Download the skill