flowtriq（flowtriq 搜索）

Flowtriq Monitor Skill

This skill lets you monitor your Flowtriq-protected infrastructure directly from your OpenClaw agent. Ask natural language questions and the agent will query the Flowtriq API and interpret the results for you.

Base URL: https://flowtriq.com Auth headers required on all agent endpoints:

Authorization: Bearer $FLOWTRIQ_API_KEY
X-Node-UUID: $FLOWTRIQ_NODE_UUID

What You Can Ask

• "Are any of my nodes under attack right now?"
• "What's the current traffic on my server?"
• "Show me recent incidents"
• "What was the peak PPS in the last attack?"
• "Is my agent online?"
• "Pull my node config and thresholds"
• "Any IOC matches recently?"

API Calls Reference

Check Agent Health / Node Status

GET https://flowtriq.com/api/health

No auth needed. Returns { "status": "ok" }. Use this first to confirm the Flowtriq platform is reachable before making authenticated calls.

Get Node Config + Current Thresholds

GET https://flowtriq.com/api/v1/agent/config
Headers: Authorization + X-Node-UUID

Returns the node's current PPS threshold, baseline stats (p99, mean), loaded IOC patterns, and any pending commands queued for the agent.

Key fields to surface to the user:

• pps_threshold — what PPS level triggers an incident
• baseline.p99_pps — the node's normal p99 traffic level
• baseline.mean_pps — average baseline
• ioc_patterns — count how many are loaded
• pending_commands — flag any that exist, user may want to review

Get Latest Metrics (Current Traffic)

POST https://flowtriq.com/api/v1/agent/metrics
Headers: Authorization + X-Node-UUID

The agent POSTs metrics to this endpoint every second. To read current traffic state, use the dashboard API. Since the dashboard endpoints require session auth, instruct the user to check https://flowtriq.com/dashboard for a live view, or pull from config endpoint for baseline context.

Submit / Check Incidents

POST https://flowtriq.com/api/v1/agent/incident
Headers: Authorization + X-Node-UUID

The agent uses this to open and update incidents. When interpreting incident data returned or confirmed by the user, extract and explain:

Attack family plain-English map:

• udp_flood = UDP Flood
• syn_flood = SYN Flood
• http_flood = HTTP Flood
• icmp_flood = ICMP/Ping Flood
• dns_flood = DNS Amplification
• multi_vector = Multi-Vector (combined attack types)
• unknown = Unclassified

Severity guidance:

• low = below 2x baseline, informational
• medium = 2-5x baseline, monitor closely
• high = 5-20x baseline, mitigation likely active
• critical = 20x+ baseline, full response mode

Register a New Node (if user asks)

POST https://flowtriq.com/api/deploy
Headers: Authorization: Bearer YOUR_DEPLOY_TOKEN
Body: { "name": "node-name", "ip": "x.x.x.x" }

Note: this uses a deploy token not the node API key. The user must provide their deploy token separately. Returns a new node_uuid and api_key for the registered node.

Interpreting Results for the User

Always translate raw API data into plain NOC-style summaries. Example patterns:

All clear:

"Your node is online. Baseline p99 is 12,000 PPS, threshold set at 50,000 PPS. No active incidents. 8 IOC patterns loaded."

Active attack:

"Active incident on [node]. SYN flood, critical severity. Peaked at 47,821 PPS (1.7 Gbps). 3,241 source IPs, spoofing detected. IOC match: mirai-variant (94% confidence). Mitigation should be active — check your dashboard for FlowSpec rule status."

Degraded / agent offline:

"The Flowtriq platform is reachable but your node hasn't sent a heartbeat recently. Your agent may be down. Run sudo systemctl status ftagent on the server to check."

Error Handling

Setup Reminder (if user hasn't configured env vars)

If FLOWTRIQ_API_KEY or FLOWTRIQ_NODE_UUID are missing, tell the user:

"To use this skill you need two environment variables set in OpenClaw: - FLOWTRIQ_API_KEY — find this in your Flowtriq dashboard under Settings > API Keys - FLOWTRIQ_NODE_UUID — find this on your node's settings page Add them to your OpenClaw env config and reload the agent."

Reference Files

• references/attack-types.md — detailed breakdown of all 8 attack

families, what causes them, and recommended responses

• references/api-endpoints.md — full endpoint reference with all

parameters for quick lookup during complex tasks