Claude Recall is a long-term memory layer for MoltBot and Claude Code that automatically learns and recalls project context. It benefits operations teams by maintaining continuity across interactions, reducing repetitive explanations, and improving efficiency. The skill integrates with Claude Code and MoltBot, enhancing workflows by retaining context from previous interactions.
git clone https://github.com/nhevers/claude-recall.gitClaude Recall adds persistent memory to Claude Code and MoltBot, automatically capturing discoveries, decisions, and code patterns from each session. At the start of new sessions, it injects relevant context from past work, eliminating the need to re-explain project details. The system uses semantic search via MCP tools to find contextual information, tracks token usage and session analytics, and provides a web viewer at localhost:37777 for browsing and organizing memories. Teams working on ongoing projects benefit from reduced context-switching overhead and faster onboarding to previous work. Memory is stored locally in SQLite and searchable through ChromaDB vector search.
[{"step":"Identify the task or context you need to retrieve. Replace [TASK] in the prompt template with your specific goal (e.g., 'debugging the payment processing module' or 'onboarding a new team member to the data pipeline project').","tip":"Be as specific as possible about the aspects you need recalled (e.g., decisions, dependencies, code snippets) to get the most relevant results."},{"step":"Open Claude Code or MoltBot and paste the customized prompt. Ensure you have enabled Claude Recall for your project (check the settings in your workspace).","tip":"If you're working in a new project, run a quick test by asking Claude Recall to summarize the last 3 interactions to verify it's capturing context correctly."},{"step":"Review the recalled context and highlight any gaps or inaccuracies. Use the output to inform your next steps, such as updating documentation or scheduling a sync with the team.","tip":"If the context is incomplete, explicitly ask Claude Recall to search for additional details (e.g., 'Retrieve all discussions about the payment module from the past 3 months')."},{"step":"After completing the task, update the project context in Claude Recall by summarizing key decisions, changes, or new dependencies. This ensures future interactions benefit from the latest information.","tip":"Use the format: 'Key updates for [PROJECT]: [SUMMARY].' to keep the recall layer concise and actionable."}]
Maintaining auth implementation details across development sessions
Tracking project decisions and architectural choices over time
Reducing repetitive context explanations in multi-session workflows
Building institutional knowledge for operations and engineering teams
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/nhevers/claude-recallCopy 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 Claude Recall to retrieve and synthesize the most relevant project context for [TASK]. Focus on [SPECIFIC_ASPECTS] such as [EXAMPLE_1], [EXAMPLE_2], or [EXAMPLE_3]. Provide a concise summary of the key details, decisions, and dependencies that impact this task. If no relevant context exists, state that explicitly and suggest how to capture it going forward.
For the task of refactoring the `user-authentication` service in the AcmeCorp platform, Claude Recall retrieved the following context: **Key Decisions:** - The team decided in Q2 2024 to migrate from JWT to OAuth2 for all authentication flows, with a 6-month sunset period for legacy tokens. This was documented in Slack thread #auth-2024-06-15. - The legacy `password-reset` endpoint was deprecated in favor of a new `/v2/reset-password` endpoint, which now requires email verification via SendGrid. This change was implemented in PR #421 but never fully communicated to the frontend team. **Dependencies:** - The new OAuth2 implementation depends on the `auth-service` v3.2.1, which is currently in staging but not yet deployed to production. - The frontend team is still using the deprecated `/reset-password` endpoint, causing user confusion during password resets. **Open Questions:** - Should we enforce the OAuth2 migration deadline despite the frontend team's delay? - How should we handle tokens issued during the transition period? **Recommendation:** Prioritize a sync with the frontend team to align on the OAuth2 migration timeline and communicate the deprecation of the old endpoint. The refactoring task should include a backward-compatible fallback for legacy tokens until the migration is complete.
Automate your browser workflows effortlessly
AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
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