Automated summary
Memory Kit persists Claude Code session state (goals, decisions, patterns, open questions) into a MEMORY.md file at the project root so it survives context compaction and session restarts.
Strengths
- Covers the full memory lifecycle with five dedicated commands: save, load, update, share and audit
- Writes a structured MEMORY.md with timestamped session metadata, rationale-backed decisions and unresolved questions
- The audit command flags stale entries (older than 7 days or tied to finished work) before prompting for removal
- Git-based sharing lets several teammates load the same project context
Limitations
- Relies entirely on a single MEMORY.md file at the project root: no per-branch memory or splitting by topic
- Sharing requires a git repo with push access, and auditing needs manual user confirmation for every removal
- Entries are generated from conversation context, so a session that was already compacted may yield an incomplete snapshot
Use it when you run long or recurring Claude Code sessions on the same project and want to avoid restating decisions and goals after each compaction.
Skip it for one-off tasks or throwaway scripts, where a versioned MEMORY.md would add clutter without real benefit.
Security analysis
CautionThe skill is not overtly malicious—it reads and writes a local memory file and uses git for sharing—but it executes shell status commands at load, grants broad git access, and can push potentially sensitive session context to a remote. These are legitimate but powerful capabilities, so it warrants caution rather than a clean safe rating.
- •The skill auto-executes shell snippets via `!` at load time (`[ -f MEMORY.md ]`, `wc`, `date -r`); the commands are benign but establish command execution outside the declared `Bash(git:*)` scope.
- •`/memory-share` commits and pushes `MEMORY.md` to a git remote; session memory can contain sensitive project details, so sharing can leak context if invoked without review.
- •`Bash(git:*)` is broader than needed for commit/push and could permit destructive git operations such as reset, clean, or force-push if the model chooses them.
- •`MEMORY.md` is loaded as persistent context and may be shared across teammates; a compromised or malicious file can act as a prompt-injection vector.
Examples
Save my memory before the context gets compacted.Load memory and tell me where we left off — should we resume the previous context or start fresh?Share memory with the team via git, then run a memory audit to prune entries older than 7 days.name: memory-kit description: 'Persistent context management for Claude Code sessions. Save, load, update,
share, and audit session memory via MEMORY.md. Prevents context loss on
compaction or session restart. Use when starting a session, before compaction,
syncing context across teammates, or pruning stale memory entries.
Trigger with "save memory", "load memory", "memory audit", "memory share".
' allowed-tools: Read, Write, Edit, Bash(git:*) version: 1.1.0 author: builtbyzac license: MIT tags:
- productivity
- audit
- memory-kit compatibility: Designed for Claude Code
Memory Kit
Current State
![ -f MEMORY.md ] && echo "MEMORY.md: $(wc -l < MEMORY.md) lines, last modified $(date -r MEMORY.md '+%Y-%m-%d %H:%M')" || echo "No MEMORY.md found"
![ -f tasks/current-task.md ] && echo "Active task file found" || echo "No task file"
Overview
Claude Code sessions lose context on compaction and restart. Memory Kit persists
session state (goals, decisions, patterns, open questions) to a MEMORY.md file
that survives across sessions.
Five commands cover the full lifecycle:
/memory-save— snapshot before compaction/memory-load— restore at session start/memory-update— log a decision mid-session/memory-share— push to git for teammates/memory-audit— prune stale entries
Prerequisites
- A git repository (for
/memory-share) - Write access to the project root (MEMORY.md lives there)
Instructions
- On session start — check for existing
MEMORY.mdin project root. If found, read and summarize the saved state. Ask the user whether to resume previous context or start fresh. - On save (
/memory-save) — scan the current conversation for goals, decisions, patterns, and open questions. Write a structured snapshot toMEMORY.mdwith timestamped sections. - On update (
/memory-update) — append the user's decision or note to the appropriate section inMEMORY.mdwithout overwriting existing content. - On share (
/memory-share) — commitMEMORY.mdand push to the remote branch so teammates can load the same context. - On audit (
/memory-audit) — review all entries inMEMORY.md, flag stale items (older than 7 days or referencing completed work), and prompt the user to confirm removal.
Output
The skill produces and maintains a MEMORY.md file containing:
- Session metadata: Timestamp, branch, and project name
- Goals: Current objectives carried across sessions
- Decisions: Key choices made with rationale
- Patterns: Recurring approaches or conventions discovered
- Open questions: Unresolved items requiring future attention
Output Format
For the MEMORY.md template structure, see output-format.md.
Error Handling
For error scenarios and recovery behavior, see error-handling.md.
Examples
Save before compaction:
"Save my memory" → reads current context, writes snapshot to MEMORY.md
Load at session start:
"Load memory" → reads MEMORY.md, summarizes state, asks to resume or start new
Quick mid-session log:
"Log decision: using Postgres over SQLite for concurrent writes" → appends to Decisions section
Team sync:
"Share memory" → runs
scripts/memory-share.sh, confirms push
Resources
- output-format.md — MEMORY.md template structure and section schema
- error-handling.md — error scenarios, recovery behavior, and edge cases
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/memory-kit-for-claude-code-mrwakclk' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/memory-kit-for-claude-code-mrwakclk/download' --output '.skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md.part' && mv '.skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md.part' '.skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
Environment declared by the author: Designed for Claude Code
Tools declared by the author (check availability in your agent): Read, Write, Edit, Bash(git:*)
Original excerpt: Prerequisites
- A git repository (for
/memory-share) - Write access to the project root (MEMORY.md lives there)
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.
Claude Code
test -f '.skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md' && test -s '.skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/memory-kit' && cp -R '.skill-downloads/memory-kit-for-claude-code-mrwakclk' '.claude/skills/memory-kit'OpenSkills
OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.
Official OpenSkills installationopenskills install './.skill-downloads/memory-kit-for-claude-code-mrwakclk'openskills read 'memory-kit-for-claude-code-mrwakclk'Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.
For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.
openskills syncManual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/memory-kit-for-claude-code-mrwakclk/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
PDF Reader
Productivity
Read and extract text from PDF files, including documents, reports, contracts, and spreadsheets. Works with local files, URLs, and WhatsApp attachments. Use whenever PDF content is needed, even if not explicitly requested.
Discord Messaging Workflows
Productivity
Use the message tool to send, edit, and moderate Discord messages with interactive components.
Interactive Verdict Preview
Productivity
Generate an interactive HTML preview of keep/drop verdicts for maintainer review, then process the pasted feedback JSON.