cogops

name	cogops
description	CogOps Swarm Intelligence — query live swarm state, trigger narrations, inject signals

CogOps Swarm Intelligence Skill

You have access to a living swarm intelligence organism called CogOps. It runs 50,000+ agents in Rust/WASM with real SIRS epidemiology, Darwinian evolution (6 heritable genes), spatial hash grids, 6-channel pheromone fields, and Ebbinghaus memory decay. The swarm ingests live crypto prices (BTC, ETH, SOL from CoinGecko) and GitHub developer activity as sensory input.

Available Tools

Query Swarm State

Fetch the current state of the swarm organism including all metrics, alerts, and real-world data feeds.

curl -s http://localhost:3000/api/swarm | python3 -m json.tool

Returns: agent count, mean surprise, mean health, R₀ effective/base, gene diversity, generation, alerts with LLM analysis.

Query Real-World Data Feeds

Get the live crypto and GitHub data feeding the swarm.

# Crypto prices (BTC, ETH, SOL with 24h change)
curl -s http://localhost:3000/api/feeds/crypto | python3 -m json.tool

# GitHub activity (event counts, activity score)
curl -s http://localhost:3000/api/feeds/github | python3 -m json.tool

Request Narration

Ask Gemini to explain what the swarm is currently sensing. Provide the event type and current metrics.

curl -s -X POST http://localhost:3000/api/swarm/narrate \
  -H "Content-Type: application/json" \
  -d '{
    "eventType": "surprise_cascade",
    "metrics": {
      "tick": 1000,
      "nAgents": 50000,
      "meanSurprise": 0.18,
      "meanHealth": 0.72,
      "r0Eff": 1.35,
      "r0Base": 0.8,
      "geneDiversity": 0.042,
      "meanGeneration": 5.2,
      "surprisedCount": 8500,
      "peakSurprise": 0.95
    },
    "prevMetrics": {
      "meanSurprise": 0.05,
      "r0Eff": 0.7
    },
    "realWorld": {
      "crypto": {
        "bitcoin": {"usd": 65000, "usd_24h_change": -3.5},
        "ethereum": {"usd": 1900, "usd_24h_change": -5.1},
        "solana": {"usd": 80, "usd_24h_change": -6.2}
      }
    }
  }' | python3 -m json.tool

Returns: A Gemini-generated plain-English explanation of what the swarm is experiencing, connected to real-world data.

Event types: surprise_cascade, r0_critical, gene_collapse, health_crisis, mass_death

Inject Signal Into Swarm

Inject an environmental shock into the swarm. Use this when the user wants to test the swarm's reaction to a scenario.

curl -s -X POST http://localhost:3000/api/openclaw/inject \
  -H "Content-Type: application/json" \
  -d '{
    "x": 500,
    "y": 500,
    "radius": 100,
    "intensity": 0.8,
    "reason": "User requested Bitcoin crash simulation"
  }' | python3 -m json.tool

Parameters:

• x, y: position in the swarm world (0-1000)
• radius: shock radius (10-200)
• intensity: surprise intensity (0.0-1.0)
• reason: human-readable reason for the injection

How to Use This Skill

When the user asks about markets, anomalies, or swarm behavior:

1. First query /api/feeds/crypto to get current prices
2. Then query /api/swarm to get the swarm's biological response
3. If interesting, request a narration via /api/swarm/narrate to get Gemini's analysis
4. Synthesize the real-world data + swarm response + narration into a clear answer

When the user says things like "what does the swarm think about Bitcoin?" or "is the market doing anything weird?", use this workflow.

Example Interactions

User: "What is the swarm sensing right now?" → Query state + crypto, narrate if R₀ > 1.0 or surprise elevated

User: "Shock the swarm — simulate a flash crash" → Inject signal with high intensity, wait 2s, query state again, narrate the response

User: "How are markers doing?" → Query crypto feed, explain prices, query swarm state to see if the organism is reacting