- Home
- Skills
- Agents & Orchestration
- Mind-Palace: Persistent Project Memory for Claude
Works with the AI Agents you already use
Mind-Palace: Persistent Project Memory for Claude
Maintain lightweight, durable project memory for Claude Code to persist context across sessions and context clears.
$9.99
Mind-Palace: Persistent Project Memory for Claude
Example session with this skill installed
I've finished the auth refactor and verified it with the test suite. Update the project memory and tell me what's next.
- Read your context and instructions
- Compiled the mind-palace
I have updated the project memory to reflect the completion of the authentication refactor. I verified the login and registration flows using the existing Vitest suite.
Project Memory Update
| Section | Status | Details |
|---|---|---|
| Current Goal | Completed | Refactor JWT middleware to use the new AuthService provider. |
| Verified Commands | Added | npm run test:auth - Runs specialized auth integration tests. |
| Key Decisions | Updated | Switched from local storage to HTTP-only cookies for session persistence. |
| Coding Conventions | Updated | All auth-related logic must reside in src/lib/auth. |
| Task Status | DONE | Auth refactor, Vitest integration, Middleware cleanup. |
| Next Steps | TODO | Implement password reset flow; Audit session timeout logic. |
Updated: Project snapshot, Key decisions, Task status
Verified: npm run test:auth (All 12 tests passed)
Next: Password reset flow implementation
Memory file: PROJECT_MEMORY.md
Next steps
- Review the new
src/lib/auth/session.tsfor edge cases in cookie expiration. - Trigger the password reset implementation using the
AuthServicetemplate.
Connects securely to your tools. The creator never sees your data.
About this skill
Losing context between Claude Code sessions or after a /clear command wastes time and tokens. This skill maintains a lightweight, durable Markdown file that captures architectural decisions, verified commands, and task progress so you can resume work immediately.
What it does
- Persistent memory management maintains a compact
PROJECT_MEMORY.mdfile to bridge context gaps between sessions. - Decision tracking records architectural rationale and coding conventions as they are discovered.
- Verified command storage saves tested terminal commands and scripts to prevent repeated trial and error.
- Task synchronization tracks progress, blockers, and next steps to provide a clear starting point for future work.
- State verification cross-references stored memory against the current repository state to ensure accuracy.
How it works
- Initialize by reading
CLAUDE.mdand the existing memory file to sync with the current project state. - Update the memory file at meaningful checkpoints, such as after a milestone is reached or a convention is established.
- Verify that any recorded facts, commands, or test results match the actual repository manifests and code.
- Compact context before session end by summarizing changes and identifying the specific next steps for the next session.
Frameworks & tools
Works with any project stack using Claude Code. It targets Markdown files like PROJECT_MEMORY.md and CLAUDE.md.
Why this beats prompting it yourself
Manual prompting often leads to context bloat or "hallucinated" project states after a long session. This skill enforces a structured workflow that separates verified facts from hypotheses and ensures durable knowledge survives a context flush.
Use cases
- Resuming a complex refactor after a weekend or a
/clearcommand. - Onboarding a new agent instance to a repository with specific, non-standard build commands.
- Keeping track of "why" a specific architectural choice was made during a fast-paced coding session.
- Managing a handoff between different sub-tasks without losing the current execution state.
Known limitations
Does not independently guarantee that a host application automatically loads the file without proper CLAUDE.md instructions. Never stores secrets or sensitive credentials.
Requires explicit project instructions to ensure host applications load the file. Does not store secrets or API keys. Memory file must be manually pruned if it grows too large.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 20 days ago