agentleFS
yigityildiz0yigityildiz0/super-memory/skills/claude-code/supermemory/SKILL.md

supermemory

Skill2 starsChanged 5 months agoMITView on GitHub

Always-on project memory skill for Claude Code, Codex, ChatGPT, and other AI coding sessions. Use automatically at the start of project or coding work, before planning, to locate or create `.ai-handoff/SUPER_MEMORY.md`, read the current packet, compare agent cursors, and keep the file updated after meaningful progress. Also use for `/supermemory`, `supermemory search`, `supermemory update`, `supermemory new`, `supermemory compress`, `supermemory delete`, `supermemory status`, Super Memory, resume, continue, handoff, context window full, limit reached, switch to Codex, switch to Claude Code, new chat continuation, deleted chat recovery, teammate handoff, devam et, kaldığım yerden devam et, limit bitti, context doldu, and project memory requests.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. supermemoryai/supermemory★ 32ksame name
    2. supermemoryai/skills★ 113same name
    3. Venkatganeshreddy/claude--skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

Show its rules and text

We split this skill into the rules it gives an agent

21 rules. 0 are common, 16 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 16

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill as a project-local memory system.

    line 8
  2. Very specific:

    If it exists, read Current Packet first.

    line 18
  3. Distinctive:

    Read only timeline entries newer than this agent's cursor when possible.

    line 19
  4. Distinctive:

    If it is missing, create it before substantial work.

    line 20
  5. Very specific:

    If older .ai-handoff/SUPERMEMORY.md or .ai-handoff/HANDOFF.md exists, migrate useful facts into SUPER_MEMORY.md.

    line 21
  6. Distinctive:

    If the root is ambiguous, ask where the project memory should live.

    line 22
  7. Distinctive:

    Do this before planning when the user asks to continue, resume, switch agents, recover from a limit, recover from a deleted chat, or work …

    line 24
  8. Distinctive:

    Use one project-local memory folder:

    line 28
  9. Distinctive:

    Do not create multiple competing memory files for the same project.

    line 33
  10. Distinctive:

    Treat these as slash-style commands:

    line 37
  11. Distinctive:

    Never save hidden reasoning, full transcripts, secrets, API keys, credentials, huge logs, or unrelated personal data.

    line 61
  12. Very specific:

    Keep SUPER_MEMORY.md compact and readable:

    line 65
  13. Distinctive:

    When Claude Code resumes:

    line 98
  14. Distinctive:

    Read timeline entries after that checkpoint.

    line 102
  15. Very specific:

    Update Claude Code last-read after absorbing them.

    line 104
  16. Distinctive:

    If the cursor is missing, read the latest packet plus the last 10 timeline entries.

    line 106
  17. Distinctive:

    Keep memory dense from the start.

    line 110
  18. Distinctive:

    Keep the latest useful timeline entries.

    line 124
  19. Distinctive:

    Remove obsolete detail unless it prevents repeating a mistake.

    line 126
  20. Distinctive:

    before ending a long turn

    line 139
  21. Very specific:

    Use scripts/supermemory.py for deterministic file handling:

    line 143

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Super Memory

Use this skill as a project-local memory system. It helps Claude Code, Codex, ChatGPT, and other agents continue from the same project state without relying on private chat history.

Credit: Super Memory by Yigit Yildiz, GitHub @yigityildiz0.

First Move

For any real project task:

  1. Identify the active project root.
  2. Look for .ai-handoff/SUPER_MEMORY.md.
  3. If it exists, read Current Packet first.
  4. Read only timeline entries newer than this agent's cursor when possible.
  5. If it is missing, create it before substantial work.
  6. If older .ai-handoff/SUPERMEMORY.md or .ai-handoff/HANDOFF.md exists, migrate useful facts into SUPER_MEMORY.md.
  7. If the root is ambiguous, ask where the project memory should live.

Do this before planning when the user asks to continue, resume, switch agents, recover from a limit, recover from a deleted chat, or work on project files.

Storage

Use one project-local memory folder:

  • .ai-handoff/SUPER_MEMORY.md: canonical current project memory.
  • .ai-handoff/LOG.md: compact older milestones.

Do not create multiple competing memory files for the same project.

Commands

Treat these as slash-style commands:

  • /supermemory search: find/read memory, absorb current packet, then read only new entries since this agent's cursor.
  • /supermemory update: refresh current packet and append a compact checkpoint after meaningful progress.
  • /supermemory new: create memory only when none exists; if one exists, update it instead of creating a duplicate.
  • /supermemory compress: preserve active state, shorten stale timeline, move old detail to LOG.md.
  • /supermemory delete: delete only .ai-handoff/ after explicit user intent and root confirmation.
  • /supermemory status: report memory path, age, latest checkpoint, and next action.

What To Save

Save operational facts:

  • goal and scope
  • project root
  • user constraints that affect the work
  • active files and why they matter
  • completed changes
  • decisions with evidence
  • commands, tests, checks, and results
  • blockers, failed attempts, and useful error text
  • next 3-7 actions
  • notes for the next agent

Never save hidden reasoning, full transcripts, secrets, API keys, credentials, huge logs, or unrelated personal data.

Format

Keep SUPER_MEMORY.md compact and readable:

# Super Memory

Last updated: <ISO datetime>
Updated by: <Codex | Claude Code | ChatGPT | Other>
Project root: <path>
Mode: compact

## Current Packet
- Goal: <1-3 bullets max>
- State: <current state>
- Active files: `<path>` = <why it matters>
- Decisions: <decision> -> <reason/evidence>
- Verification: `<command/check>` -> <result>
- Blockers: <none or current blocker>
- Next: <3-7 concrete actions>

## Agent Cursors
- Codex last-read: <checkpoint id or none>
- Claude Code last-read: <checkpoint id or none>
- ChatGPT last-read: <checkpoint id or none>
- Other last-read: <checkpoint id or none>

## Timeline
- cp-YYYYMMDDTHHMMSSffffffZ | <Agent> | <dense update>

Cursor Rule

The current packet is always relevant. Timeline entries are incremental.

When Claude Code resumes:

  1. Read Current Packet.
  2. Find Claude Code last-read.
  3. Read timeline entries after that checkpoint.
  4. Prioritize entries written by other agents after the cursor.
  5. Update Claude Code last-read after absorbing them.

If the cursor is missing, read the latest packet plus the last 10 timeline entries.

Compression Rule

Keep memory dense from the start. Prefer this style:

  • Auth fix done; changed src/auth.ts; npm test pass; next deploy.
  • Build fail: missing env VITE_API_URL; no code issue found.

Avoid this style:

  • I worked on the project and made some progress.
  • pasted conversations
  • repeated explanations

When compressing:

  1. Preserve goal, active files, decisions, verification, blockers, and next actions.
  2. Keep the latest useful timeline entries.
  3. Move old entries into LOG.md as short milestones.
  4. Remove obsolete detail unless it prevents repeating a mistake.

Update Timing

Update memory after:

  • behavior-changing edits
  • new or resolved blockers
  • changed plan or scope
  • tests/build/lint checks
  • failed attempts worth remembering
  • agent switch
  • likely context reset or limit
  • before ending a long turn

Helper Script

Use scripts/supermemory.py for deterministic file handling:

python scripts/supermemory.py search --root . --agent "Claude Code"
python scripts/supermemory.py update --root . --agent "Claude Code" --message "Implemented X; tests passed."
python scripts/supermemory.py new --root . --agent "Claude Code"
python scripts/supermemory.py compress --root . --keep 20
python scripts/supermemory.py status --root .
python scripts/supermemory.py delete --root . --yes

Manual edits are acceptable when they produce a clearer current packet.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.