skills.rest
lurodrisilva
lurodrisilvaCOMMUNITY
Explore all skills by this author →
GitHub

azure-sre-agent

Configure and govern Azure SRE Agent extension primitives, MCP connectors, and permission gates.

Updated Apr 11, 2026
Search Tags:#root cause analysis#azure#remediation#SRE#permission gate#incident triage#MCP connectors
One-click install
npx skills add https://github.com/lurodrisilva/personal-skills --skill azure-sre-agent-lurodrisilva
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: azure-sre-agent
Source: https://github.com/lurodrisilva/personal-skills/tree/main/operations/azure-sre-agent
Command: npx skills add https://github.com/lurodrisilva/personal-skills --skill azure-sre-agent-lurodrisilva

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Operating Microsoft's Azure SRE Agent (Preview) safely in production requires knowing its extension model, MCP connector constraints, and approval doctrine — without this guidance, teams risk unbounded blast radius, blown tool budgets, and un-audited autonomous remediation. ## Core Features & Use Cases - Extension Model Guidance: Covers the six extension primitives — Skills/runbooks, built-in Subagents, Python tools, MCP servers, agent hooks, and the Permission gate — with rules for choosing the lightest fitting primitive. - MCP Connector Model: Documents Streamable-HTTP vs stdio transports, Bearer/custom-headers/managed-identity auth, namespaced connection-id_tool registration, and the hard 80-tool-per-agent budget. - Approval Doctrine: Encodes the propose-then-approve safety model where the Permission gate evaluates every tool call (approve/policy/block) and composes with Managed-Identity RBAC, with audit telemetry routed to your own Application Insights. - Use Case: When wiring a Dynatrace MCP connector into Azure SRE Agent for AKS incident triage, use this Skill to scope tools per subagent with mcp_tools:, define gate rules before granting mutating tools, and keep auto-remediation opt-in. ## Quick Start Ask the AI to help configure an Azure SRE Agent MCP connector with proper permission-gate rules and a scoped tool budget for incident triage on AKS.

Frequently Asked Questions about azure-sre-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add an MCP server connector to Azure SRE Agent?▼

Add MCP connectors via Builder → Connectors → + Add connector → MCP Server. Remote servers use Streamable-HTTP over HTTPS; local stdio servers run on Node 20, Python 3.12, or .NET 9 with no Docker. Tools register namespaced as connection-id_toolname.

What is the Azure SRE Agent permission gate?▼

The Permission gate is a pre-execution safety layer that evaluates every proposed tool call and can approve, enforce policy, or block it. It composes with Azure RBAC on the agent's Managed Identity, and all decisions should be audited to your own Application Insights.

Does Azure SRE Agent auto-remediate incidents without approval?▼

No, Azure SRE Agent proposes mitigations but does not apply them without human approval by default. Auto-approve should only be enabled for proven-safe, reversible, tightly-scoped action classes, loosening one dimension at a time.

What is the tool limit for Azure SRE Agent MCP connectors?▼

Azure SRE Agent enforces a hard limit of 80 tools per agent, counting native and MCP tools combined. Scope connectors per subagent using mcp_tools: wildcards rather than importing a connector's full toolset.

Which auth methods do Azure SRE Agent MCP connectors support?▼

MCP connectors support Bearer tokens for most SaaS servers like GitHub and Dynatrace, custom headers for services like Datadog, and managed identity for Azure services over stdio. Partner connectors ship with locked auth and prefilled URLs.

Is Azure SRE Agent generally available and version-pinned?▼

No, Azure SRE Agent is a Preview managed service, so no version should be pinned and surfaces may change. Verify every primitive and limit against Microsoft Learn at learn.microsoft.com/azure/sre-agent before relying on it.

Related Skills

Explore More Software Engineering Skills→
jorgeasaurusjorgeasaurus
12

azure-rbac

Determine Azure RBAC roles and generate CLI and Bicep assignment code.

Community
Intermediate
Zhachory1Zhachory1

human-approval-gate

Prepares decision packages with evidence, risks, and rollback plans for human signoff.

Community
Intermediate
Happy-Technologies-LLCHappy-Technologies-LLC
34

observability-integration

Automate observability data integration from Datadog, Dynatrace, and New Relic into ServiceNow ITOM.

Official
Advanced
theslashdojotheslashdojo

kubectl

Automates Kubernetes cluster management via natural-language-driven tool orchestration across hybrid and on-premise environments, enabling automated workflows, rapid prototyping, and hands-free operation for complex tasks requiring minimal manual effortpping, and quick-turnaround CI/CD pipelines with automated error recovery and built-in safeguards for unattended execution.

Official
Advanced