- Home
- Skills
- Agents & Orchestration
- neurodynamic brain
More screenshots
Works with the AI Agents you already use
neurodynamic brain
A file-based, cross-project memory system for coding agents with selective recall and durable lesson persistence.
$9
neurodynamic brain
Example session with this skill installed
I just spent an hour debugging the CI pipeline. The issue was that the setup-node action needs cache: 'npm' explicitly set in our self-hosted runners. Remember this for next time.
- Read your context and instructions
- Compiled the neurodynamic brain
I have recorded this lesson in the neurodynamic brain. I allocated a new cell, verified it against the tag vocabulary, and updated the index.
Knowledge Cell Created: id-042
| Field | Value |
|---|---|
| ID | 042 |
| Slug | ci-npm-cache-fix |
| Tags | @ci, @npm, @github-actions |
| Scope | global (self-hosted runners) |
| Hook | Explicit npm caching is required on self-hosted runners to prevent CI timeouts. |
Cell Body:
Found that actions/setup-node fails to resolve paths correctly on our custom runner images unless cache: 'npm' is explicitly defined. Without it, the runner attempts to scan the entire root directory, leading to job timeouts.
Why: Default caching logic assumes standard GitHub-hosted filesystem paths.
How to apply: Always include cache: 'npm' in the with block of the setup-node step for any new workflow files.
Next steps
- Run
ndbrain doctorto confirm the new cell is correctly indexed. - The next time you ask to create a CI workflow, I will trigger
ndbrain recall @ciand retrieve this requirement automatically. - Consider running
ndbrain handoffif you are finished with this session to sync this update to your other environments.
Connects securely to your tools. The creator never sees your data.
About this skill
Coding agents often suffer from a short-term memory that resets every session, losing hard-won lessons about your specific architectural quirks or edge cases. This skill implements a file-based, cross-project memory system that allows agents to recall facts, handle machine handoffs, and persist lessons across different repositories without context window bloat.
What it does
- Selective Recall uses a tag-based index to load only the specific facts relevant to the current task, keeping context windows clean.
- Knowledge Persistence creates durable Markdown cells for lessons that should survive session resets or repository switches.
- Handoff Management generates structured
HANDOFF.mdupdates to resume complex tasks across different machines or days. - Health Monitoring provides a
doctorandscorecommand to identify stale information, duplicate entries, or structural errors. - Gate Protocol enforces a six-stage workflow (Intent, Recall, Scope, Build, Verify, Learn) to ensure every task uses existing knowledge.
How it works
- Initialize the memory directory using
ndbrain initto set up the global storage and link the protocol to your agent file. - Recall relevant context at the start of a task by running
ndbrain recall <tags>to search the central index. - Reason through the task using the fetched knowledge cells while following the G0-G5 gate loop.
- Write-back new durable lessons using
ndbrain newto create a new indexed cell once a problem is solved.
Frameworks & tools
Built as a zero-dependency Node.js utility (Node 18+). It integrates with Claude Code, Cursor, Codex, and Gemini via standard agent protocol files like CLAUDE.md or .cursorrules.
Why this beats prompting it yourself
Prompting alone cannot bridge the gap between two different repositories or a machine restart. This skill provides the underlying CLI and file structure to ensure your agent actually saves what it learns in a way that remains searchable and cheap to retrieve months later.
Use cases
- Multi-repo consistency for maintaining the same linting or deployment standards across different projects.
- Onboarding aid for agents to remember specific "gotchas" in a legacy codebase that aren't documented in the README.
- Session resumption when moving a complex refactor from a local machine to a remote dev environment.
Known limitations
Requires a manual sync method (like Git) to share the memory directory between machines. It is not an automated RAG system; the agent must actively trigger recall and write-back commands.
Requires Node 18+. Does not automate cloud syncing between machines. Agent must manually trigger recall/write-back commands. No built-in encryption for stored cells.
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 11 days ago