UnseveredMemory is a lightweight memory system for Claude Code. It enables persistent storage and retrieval of data across sessions. Developers and operations teams benefit from maintaining context in long-running workflows. It connects to Claude Code hooks and can be integrated into bash scripts and markdown documentation.
git clone https://github.com/blas0/UnseveredMemory.gitUnsevered Memory solves context loss in long Claude Code sessions by enforcing persistence through three hooks: SessionStart loads prior state, UserPromptSubmit injects memory reminders on every prompt (surviving context compaction), and SessionEnd archives work. Unlike simple context injection, this system maintains state across sessions using markdown files stored in .claude/memory/ and .ai/ directories, with daily archives and decision logs. It works offline, requires no databases, and integrates natively with Claude Code's hook system. Developers managing multi-session projects, DevOps automation, and complex workflows benefit from reliable memory enforcement that survives the model's natural tendency to forget earlier instructions.
["Install UnseveredMemory in your Claude Code environment using: `pip install unsevered-memory` or follow the [official setup guide](https://github.com/your-repo/unsevered-memory).","Initialize memory for your project by running: `memory init --project [PROJECT_NAME]`. This creates a dedicated storage directory in ~/.claudecode/memory/.","Save data persistently using: `memory save --key [MEMORY_KEY] --data '[YOUR_JSON_DATA]'`. Replace [MEMORY_KEY] with a unique identifier for your data (e.g., 'config_v1').","Retrieve data anytime with: `memory get --key [MEMORY_KEY]`. Use the `--summary` flag to get a human-readable summary instead of raw JSON.","Integrate with scripts by using the CLI tool or importing the Python module directly. For bash scripts, use `memory save` and `memory get` commands with variables: `DATA=$(memory get --key config_v1 --summary)`."]
Maintaining context in multi-day debugging sessions and troubleshooting workflows
Preserving architectural decisions and patterns across development sprints
Managing long-running DevOps automation tasks with cross-session state
Tracking operations logs and findings in live scratchpads during development
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/blas0/UnseveredMemoryCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Use UnseveredMemory to store and retrieve persistent data for [PROJECT_NAME]. First, save the current state of [DATA_TO_SAVE] with the key '[MEMORY_KEY]'. Then, retrieve and summarize the stored data. Finally, update the memory with [NEW_DATA] under the same key. Ensure all operations are logged in [LOG_FILE].
```markdown
# UnseveredMemory Session Log - Project: AlphaAnalytics
## Initialization (2024-05-20 14:32:18 UTC)
- Memory system initialized for Project AlphaAnalytics.
- No existing data found under key 'alpha_state_v2'.
## Data Storage (2024-05-20 14:32:45 UTC)
- Saved dataset: {"active_users": 1247, "api_calls": 8923, "error_rate": 0.02}
- Key used: 'alpha_state_v2'
- Location: ~/.claudecode/memory/alpha_state_v2.json
## Data Retrieval (2024-05-20 14:33:01 UTC)
- Retrieved data for 'alpha_state_v2':
```json
{
"active_users": 1247,
"api_calls": 8923,
"error_rate": 0.02,
"last_updated": "2024-05-20T14:32:45Z"
}
```
- Summary: The project has 1,247 active users with 8,923 API calls and a 2% error rate. No critical issues detected.
## Data Update (2024-05-20 14:34:12 UTC)
- Updated dataset: {"active_users": 1258, "api_calls": 9012, "error_rate": 0.018}
- Key used: 'alpha_state_v2'
- Changes detected: +11 users, +89 API calls, error rate improved by 0.2%
## Final State (2024-05-20 14:34:12 UTC)
- Memory successfully updated and synchronized.
- Log written to ~/.claudecode/memory/alpha_state_v2.log
```AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
Automate security compliance and monitor real-time security posture seamlessly.
Automate your spreadsheet tasks with AI power
Agentic AI Workflow platform
Connected workspace for docs, wikis, and projects
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan