Skills Framework

Turn operational knowledge into safe agent action.

Each Skill packages a task, its context, permissions and guardrails into a capability every agent can reuse.

SKILL.mdAgents

Write it once. Every agent runs it.

A Skill is an AI-native runbook. Skills let Frontier Investigation Agents work the way your team already works.

Inside every Skill.

A Skill is a self-contained, declarative unit. It defines one cloud operation: what triggers it, which tools it uses, what guardrails constrain it and how its output is structured.

  • Triggers and tools

    Natural language patterns activate the Skill. It names the CLI tools, APIs and pre-authenticated connections it needs.

  • Guardrails built in

    Safety constraints limit the scope: read-only, approval gates, no-delete.

  • Typed output, chained

    Structured output that dashboards and other Skills can consume. Chain triggers fire follow-up Skills based on the result.

From trigger to resolution.

100+ pre-built Skills across six domains: cost, security, operations, performance, sustainability and reliability.

  • Detect

    An alert fires, a user types a question in Slack, or a schedule triggers. The agent identifies which skill matches.

  • Execute

    The agent activates the skill, uses its defined tools and connections, respects guardrails, and produces structured output.

  • Act and chain

    Results are returned in a typed schema. Chain triggers automatically fire follow-up skills based on conditions.

Browse all 100+ skills in the library

Build any Skill in minutes.

Turn your runbooks, SOPs and tribal knowledge into reusable Skills. Describe what you need in plain language and CloudThinker handles the rest.

  • Describe

    Describe the operation in natural language

  • Generate

    CloudThinker generates SKILL.md with all components

  • Ship

    Test, refine and deploy. Your Skill is live.

Frontier Investigation Agents

Engineers get their time back. For creative work.

Agents investigate every issue across application, cloud, security, database and Kubernetes, find the why and propose the fix. Your engineers approve, under your policies and audit trail.