SiClaw

An AI agent platform for site reliability engineering that automates deep incident investigation. It solves the problem of slow, manual root cause analysis by using specialized agents that collect data, form hypotheses, validate them in parallel, and produce structured reports with evidence and confidence scores. The platform targets SRE and DevOps teams in enterprises, offering a read-only, open-source approach with multi-agent collaboration and integration with existing observability and alerting tools. It is delivered as a self-hosted or cloud service, with a focus on security and team collaboration.

Key features

• Multi-agent workspace with specialized domain agents
• Read-only investigation by default
• Hypothesis-driven 4-phase investigation engine
• Structured root cause reports with confidence scores
• Per-agent skill system with mandatory review
• Knowledge library for runbooks and docs
• Alert-driven channels (Slack, Discord, etc.)
• Cron patrols for scheduled health checks
• Integrations with Kubernetes, Prometheus, Grafana, etc.