RunbookAI – Hypothesis-driven incident investigation agent(open source)

Name: RunbookAI – Hypothesis-driven incident investigation agent(open source)
Availability: InStock
Author: EmTekker

by EmTekker·Feb 28, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●●●BangerSolve My ProblemWizardry

Hypothesis-pruning incident agent with approval gates beats chaos engineering explorers.

Strengths

•Hypothesis branching + causal query pruning is smarter than broad LLM exploration; avoids golden-hammer mistakes.
•Full audit trail and human approval gates for every mutation lock down safety-critical ops use cases.
•Contextual learning from runbooks/postmortems/docs prevents investigation in a vacuum; MCP integration into Claude Code is genuinely clever.

Weaknesses

•Incident response tooling is emerging but not new; CompilerLA, PagerDuty Automation, and Atlassian Opsgenie already exist.
•Early-stage confidence scoring and multi-cloud support claims need production evidence.

Post Description

RunbookAI is an open-source CLI agent that investigates production incidents using hypothesis-driven reasoning. When an alert fires, it forms ranked hypotheses, runs targeted queries against your infrastructure (AWS, Kubernetes, CloudWatch), and systematically narrows to root cause.

Key design decisions:

- Hypothesis branching/pruning instead of broad exploration. Forms 3-5 hypotheses, tests each with causal queries, prunes dead ends, branches deeper on strong evidence (max depth: 4). - Every mutation requires human approval. Full audit trail of what the agent thought, what it queried, and why. - Pulls context from your runbooks, postmortems, and architecture docs (Confluence, Notion, Google Drive, or local markdown). The agent doesn't investigate in a vacuum. - Deep Claude Code integration — auto-injects relevant operational context into coding sessions via MCP.

Try the demo without any API keys: npx @runbook-agent/runbook demo

GitHub: https://github.com/Runbook-Agent/RunbookAI

Would love to hear any questions/feedback!

Similar Projects

Infrastructure●●Solid

RunbookAI – Stop scrolling dashboards at 3 a.m., let AI investigate

The project converts on-call triage into a hypothesis-driven agent that forms and prunes hypotheses, fetches evidence from CloudWatch/Kubernetes and your runbooks, and surfaces an investigation plus approval-gated remediation steps. I like the npx demo, read-only-by-default K8s stance, and built-in audit trail; the obvious caveat is its dependence on proprietary LLM keys and the ops work needed before trusting any mutating actions in production.

Solve My ProblemNiche GemWizardry

EmTekker

103mo ago

Developer Tools●●Solid