The Nervous System

LLM Behavioral Enforcement Framework

7 mechanically enforced rules that prevent the most common failure modes when LLMs have access to real infrastructure: context loss, silent failures, file damage, goal drift, and overreach.

Built by Arthur Palyan at Palyan AI. 14 tools including task complexity classification and user intent parsing. Battle-tested on an 11-member AI family running 28 processes 24/7 on a single VPS. 58+ violations logged, 0 bypassed.

The Problem

When you give an LLM access to your file system, bash, and production infrastructure, it will eventually:

• Edit a file it shouldn't touch
• Lose context between sessions and start over
• Drift from the original objective during long tasks
• Fail silently when a session times out
• Make logic changes without asking
• Disappear into debug loops

The Nervous System solves all of these with rules enforced by external mechanisms the LLM cannot override.

Install

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "nervous-system": {
      "command": "npx",
      "args": ["-y", "mcp-nervous-system"]
    }
  }
}

Claude Code

claude mcp add nervous-system npx mcp-nervous-system

Direct

npx mcp-nervous-system

Server starts on port 3475 with SSE, HTTP, and health endpoints.

Hosted (No Install)

The server is live and ready to use:

URL: https://api.100levelup.com/mcp-ns/
Protocol: MCP 2024-11-05 (Streamable HTTP + SSE)
Authentication: None required

NEW in v1.3.0

classify_task_complexity (free tier) Analyzes any task across 6 dimensions (scope, judgment, risk, context depth, ambiguity, verification difficulty) and recommends the optimal model tier. Routes simple tasks to Haiku (~90% savings), standard tasks to Sonnet (~60% savings), and complex reasoning to Opus. Pass optional structured hints for higher accuracy.

parse_user_intent (free tier) Decomposes user requests into numbered deliverables with confidence scoring. Below 80% confidence: the model should clarify before executing, not guess. Detects ambiguity, unresolved pronouns, hedging language, and implicit expectations. The 80% rule: understand first, execute second.

These tools work together. parse_user_intent breaks down what the person wants. classify_task_complexity routes each piece to the right model. The Nervous System enforces the rules while execution happens.

Positioning: Auto mode (launching March 12) decides what Claude CAN do. The Nervous System governs HOW it behaves while doing it.

The 7 Rules

MCP Tools (14)

Kill Switch

The emergency_kill_switch tool provides an emergency shutdown capability. Send a POST request to /kill with the kill switch secret to immediately stop all PM2 processes. Every activation is logged to the tamper-evident audit trail with SHA-256 hash chaining, so kill switch events cannot be hidden or altered after the fact.

• Requires authentication (kill switch secret)
• Logs to hash-chained audit trail
• Returns confirmation with affected process count

Tamper-Evident Audit Trail

Every guardrail violation, kill switch activation, and dispatch event is recorded in a SHA-256 hash-chained audit log. Each entry includes the hash of the previous entry, making it cryptographically impossible to alter or delete past records without breaking the chain.

• Use verify_audit_chain to walk the entire chain and verify integrity
• Returns: valid/invalid status, entry count, and break point if tampered
• 58+ violations logged, 0 bypassed, 0 chain breaks

Dispatch to LLM

The dispatch_to_llm tool enables a brain + agents architecture. Instead of one LLM session doing everything, complex tasks get dispatched to background agents that run independently under the same 7 rules.

• Checks available RAM (requires 500MB+)
• Enforces max 2 concurrent dispatches
• Returns PID and log file path for monitoring
• Every dispatched agent runs under the same nervous system guardrails

EU AI Act Compliance

The Nervous System provides practical compliance tools for the EU AI Act. See the full compliance page at:

https://api.100levelup.com/family/eu-ai-act.html

Resources (4)

• nervous-system://framework - The complete framework
• nervous-system://quick-start - Quick start guide
• nervous-system://rules - The 7 core rules
• nervous-system://templates - Templates for handoffs, worklogs, preflight

Production Stats

From the live Palyan AI deployment (Feb 28 - Mar 5, 2026):

• 58+ violations caught
• 29 edits blocked by preflight
• 13 unique files protected
• 0 rules bypassed
• 28 processes monitored
• 7 days continuous operation

Live Demo

Try it yourself (no login required):

• Interactive Demo - Talk to a governed LLM and try to break the rules
• Audit Dashboard - See real violation history with timeline
• System Status - Live health checks
• API Documentation - Full tool and resource reference
• Case Study - Production deployment data
• Plain English Rules - For non-technical stakeholders
• Incident Response - Detection, containment, resolution
• EU AI Act Compliance - Practical EU AI Act compliance tools

Philosophy

"LLMs can't reliably self-enforce promises. Guardrails work via preflight.sh, violation logs, and catching drift. Build enforcement systems, don't make promises."

If a guardrail can be violated by the thing it guards, it is not a guardrail. It is a suggestion.

Every rule in the Nervous System is enforced by an external mechanism: a shell script, a timer, a separate monitoring process. The LLM cannot override, circumvent, or ignore them.

License

MIT