noise-maker is a robust alert management gateway designed for Openclaw Skills to centralize webhooks, filter noise, and correlate incidents across multiple monitoring sources.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install noise-maker
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install noise-maker using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is noise-maker: Intelligent Alert Gateway?
noise-maker serves as a sophisticated intelligence layer between noisy monitoring sources like Zabbix or Graylog and your Openclaw Skills environment. Instead of flooding agent interfaces with raw, unorganized webhooks, this skill centralizes ingestion, validates security signatures via HMAC, and applies priority-ordered rules to ensure only actionable events reach the user. It transforms fragmented data into cohesive host-based incidents, providing a clean and summarized context for the agent.
By utilizing a dedicated PostgreSQL backend and local disk buffering, noise-maker ensures that no critical event is lost during database or network downtime. This infrastructure allows Openclaw Skills to maintain high availability and reliability in alert delivery while significantly reducing token consumption by suppressing repetitive or low-priority telemetry before it reaches the LLM.
noise-maker: Intelligent Alert Gateway Use Cases
- Consolidating alerts from Zabbix, Graylog, and other sources into a single Openclaw Skills stream.
- Suppressing monitoring noise and port-scan floods to prevent agent interface clutter.
- Correlating disparate alerts from different platforms into unified, host-based incidents.
- Implementing persistent alert queues that survive Openclaw Skills downtime or network instability.
- Using LLM-driven summarization to provide human-readable insights for complex multi-event incidents.
How noise-maker: Intelligent Alert Gateway Works
- Raw webhooks arrive at the ingest endpoint via Caddy-routed HTTPS with dedicated certificates.
- The gateway validates the source using primary HMAC signatures or secondary IP allowlists.
- All raw events are stored in the PostgreSQL database with unique de-duplication keys to prevent redundant processing.
- Priority-ordered rules are executed to determine if an alert should be escalated, suppressed, deferred, or tagged.
- The system normalizes the host name and correlates the alert to an existing open incident or creates a new one.
- If escalation is required, a minimal Telegram notification is sent directly to configured targets.
- Openclaw Skills agents interact with the gateway through the noise-maker skill to triage incidents, modify rules, or review suggestions.
noise-maker: Intelligent Alert Gateway Setup
Ensure you have a PostgreSQL instance and Python environment ready. Follow these steps to integrate the skill:
# Create the dedicated database on your PostgreSQL host
createdb sysclaw_alerts
# Clone the noise-maker skill into your Openclaw directory
git clone https://github.com/example/noise-maker ~/.openclaw/skills/noise-maker
# Install the required FastAPI and database dependencies
pip install fastapi uvicorn psycopg2-binary
# Start the gateway service on the designated port
uvicorn main:app --host 0.0.0.0 --port 8090
Finally, configure your Caddy server to reverse proxy alerts.example.com to the gateway IP and update your monitoring source webhooks to point to the new ingestion URL.
noise-maker: Intelligent Alert Gateway Data Schema & Taxonomy
The skill organizes its operations through a structured PostgreSQL schema:
- sources: Configuration for webhook origins including HMAC secrets and IP allowlists.
- alerts: Storage for all received raw payloads with normalization and de-duplication constraints.
- incidents: High-level records grouping alerts by normalized host name with open/closed status.
- rules: Logic engine for defining escalation actions based on priority and conditions.
- rule_suggestions: A staging area for LLM-proposed patterns that Openclaw Skills agents can review and apply.
noise-maker: Intelligent Alert Gateway Advanced Features
- Local Disk Buffering: Automatically caches alerts to the filesystem if the database is unreachable, replaying them with a specific flag once connectivity is restored.
- Major Incident Detection: Automatically triggers a lockdown mode to suppress escalation storms and override quiet hours during high-frequency alert floods.
- LLM Concierge Integration: Utilizes LiteLLM to provide rich summarization of multi-event incidents and suggest adaptive noise-reduction rules.
- Zero-Cost Startup Hooks: Uses a before_agent_start hook to inject high-level incident status into the context of Openclaw Skills only when attention is required.
- Target Memory Caching: Persists notification targets to a local JSON cache to ensure emergency Telegram pings can be sent even if the primary database is offline.
SKILL.md
Loading
Related Openclaw Skills
S2-Boundary-Scanner
An advanced spatial perception skill that enables Embodied AI to map and monitor physical boundaries using a 9-grid topology and mmWave radar data.
Yu Minhong Mindset Perspective
An AI-powered mindset model simulating the resilient, education-driven, and crisis-transformative strategic framework of Yu Minhong.
Eigen AI Terminal
An intelligence layer providing daily-updated signals and trends across 16 areas of the AI landscape for coding agents.
Zhou Hongyi Perspective v2.0
An AI-driven strategic framework based on 360 Group founder Zhou Hongyi's philosophy of disruptive innovation and user-centric free business models.
Nex Decision Journal
A local-first decision journal and reasoning tracker designed to eliminate cognitive bias and improve judgment accuracy over time.
CLAUDE.md Auto-Discovery
An automated context discovery tool that mirrors Claude Code behavior by injecting instructions from project-specific CLAUDE.md files.
Featured*
