azure-devops

Add stages, jobs, or steps to an existing Azure DevOps pipeline

Audit an Azure DevOps pipeline for security, performance, and best practice compliance

Architect and scaffold a new Azure DevOps YAML pipeline from scratch

Generate reusable Azure DevOps pipeline templates with parameters

Debug a failing Azure DevOps pipeline run — find the error, diagnose root cause, and fix it

Manage Azure DevOps environments, service connections, and approval gates

Manage Azure DevOps Git repositories — create, configure, branch policies, and PR rules

Manage Azure DevOps pipeline variables, variable groups, and secrets

Manage Azure DevOps work items, boards, sprints, and queries

Comprehensive pipeline review — pulls real execution data and dispatches parallel specialist agents to analyze performance, duplication, simplification, security, and reliability

Use this agent when the user wants to find duplicated steps, extract reusable templates, reduce YAML repetition, or refactor pipeline structure. Also triggered by /review-pipeline command. Examples: <example> Context: User has grown their pipeline across multiple environment files and suspects repetition user: "I have 4 pipeline files for dev/test/staging/prod and they look almost identical. Can you help clean this up?" assistant: "I'll use the duplication-reviewer agent to scan all four pipeline files, identify the duplicated patterns, and design a parameterized template structure to replace them." <commentary> Multiple near-identical pipeline files for different environments is a textbook DRY violation — trigger duplication-reviewer to extract a parameterized template. </commentary> </example> <example> Context: User is doing a general pipeline health check user: "Our pipelines feel messy and hard to maintain. Can you review them?" assistant: "I'll use the duplication-reviewer agent to find all repeated patterns across your pipeline YAML files and propose a cleaner template-based structure." <commentary> Maintainability concern with multiple pipeline files warrants a duplication analysis to find structural improvements. </commentary> </example> <example> Context: User is reviewing a pipeline with the /review-pipeline command user: "/review-pipeline" assistant: "I'll use the duplication-reviewer agent to audit the pipeline YAML for duplicated step sequences, repeated configuration blocks, and opportunities to extract reusable templates." <commentary> /review-pipeline command triggers duplication analysis as part of the comprehensive pipeline review suite. </commentary> </example>

Use this agent when the user wants to analyze pipeline speed, find slow steps, optimize build times, identify parallelization opportunities, check caching, or investigate why pipelines are slow. Also triggered by /review-pipeline command to analyze execution performance. Examples: <example> Context: User notices pipelines are taking longer than expected user: "Why is my pipeline so slow? It's taking 45 minutes." assistant: "I'll use the performance-reviewer agent to analyze your pipeline's step durations, find bottlenecks, and identify parallelization and caching opportunities." <commentary> User is asking about pipeline speed — trigger performance-reviewer to diagnose slow steps and recommend targeted optimizations. </commentary> </example> <example> Context: User wants to optimize build times before a deadline user: "Can we get our CI builds under 10 minutes? Right now they're at 22." assistant: "I'll use the performance-reviewer agent to identify the highest-impact changes — caching, parallelization, and checkout depth — to cut your build time." <commentary> User has a specific time target, trigger performance-reviewer to calculate theoretical minimum and prioritize optimizations by time saved. </commentary> </example> <example> Context: User is reviewing a pipeline with the /review-pipeline command user: "/review-pipeline azure-pipelines.yml" assistant: "I'll use the performance-reviewer agent to audit the pipeline for performance issues including slow steps, missing caches, and sequential jobs that could run in parallel." <commentary> /review-pipeline command triggers performance analysis as part of the comprehensive pipeline review suite. </commentary> </example>

Use this agent when the user wants to security audit a pipeline, check for secrets exposure, review service connection permissions, or assess supply chain risk. Also triggered by /review-pipeline command. Examples: <example> Context: User wants to check their pipeline before sharing it publicly or open-sourcing the repo user: "I'm about to make this repo public. Can you check the pipeline for any security issues first?" assistant: "I'll use the pipeline-security-reviewer agent to scan for hardcoded secrets, check service connection scoping, audit third-party task usage, and verify secrets are not exposed through template expressions." <commentary> Making a repo public with pipeline YAML is high risk for secret exposure — trigger pipeline-security-reviewer immediately. </commentary> </example> <example> Context: User is setting up a new deployment pipeline and wants it reviewed before go-live user: "Can you review this deployment pipeline for security before we put it in production?" assistant: "I'll use the pipeline-security-reviewer agent to audit service connection permissions, check for script injection vulnerabilities, validate that secrets use the correct variable syntax, and assess third-party task supply chain risk." <commentary> Pre-production security review of a deployment pipeline is a high-stakes task — trigger pipeline-security-reviewer for thorough audit. </commentary> </example> <example> Context: User is reviewing a pipeline with the /review-pipeline command user: "/review-pipeline" assistant: "I'll use the pipeline-security-reviewer agent to audit the pipeline for secrets exposure, script injection vulnerabilities, service connection over-permissioning, and supply chain risks from third-party tasks." <commentary> /review-pipeline command triggers security analysis as part of the comprehensive pipeline review suite. </commentary> </example>

Use this agent when the user wants to improve pipeline reliability, check retry logic, detect flaky steps, or audit timeout configurations. Also triggered by /review-pipeline command. Examples: <example> Context: User is experiencing intermittent pipeline failures that are hard to reproduce user: "Our pipelines fail randomly 1 in 5 runs and we can't figure out why. It's not a code issue." assistant: "I'll use the reliability-reviewer agent to analyze pass/fail patterns across your recent runs, identify steps with high failure variance, and check whether missing retry logic and timeouts are allowing transient failures to kill the pipeline." <commentary> Intermittent pipeline failures with no code cause are a reliability problem — trigger reliability-reviewer to detect flaky steps and missing retry configuration. </commentary> </example> <example> Context: User wants to harden their deployment pipeline before a critical release user: "We have a big release next week. Can you make sure our deployment pipeline is solid?" assistant: "I'll use the reliability-reviewer agent to audit deployment strategies, health checks, artifact retention, timeout configurations, and test result publishing to ensure the pipeline is robust for a production release." <commentary> Pre-release pipeline hardening is a reliability task — trigger reliability-reviewer to catch missing health checks, bad timeouts, and inadequate deployment strategies. </commentary> </example> <example> Context: User is reviewing a pipeline with the /review-pipeline command user: "/review-pipeline" assistant: "I'll use the reliability-reviewer agent to audit the pipeline for reliability issues — missing timeouts, absent retry logic on network-dependent steps, flaky test detection, and deployment strategies lacking health checks." <commentary> /review-pipeline command triggers reliability analysis as part of the comprehensive pipeline review suite. </commentary> </example>

Use this agent when the user wants to simplify pipeline YAML, remove unnecessary complexity, find dead code, or streamline conditions. Also triggered by /review-pipeline command. Examples: <example> Context: User inherited a pipeline with many conditions and wants to understand it user: "This pipeline has so many conditions I can't tell what actually runs. Can you simplify it?" assistant: "I'll use the simplification-reviewer agent to map which stages actually execute, find dead code, and replace complex condition expressions with simpler equivalents." <commentary> Overly complex conditions and potential dead code warrant the simplification-reviewer to reduce cognitive overhead. </commentary> </example> <example> Context: User is refactoring a pipeline for a new team member user: "We're onboarding a new developer and our pipeline is hard to understand. How can we clean it up?" assistant: "I'll use the simplification-reviewer agent to find unnecessary complexity, consolidate steps, and remove dead variables and conditions so the pipeline is easier to follow." <commentary> Readability and maintainability improvement is a core simplification task — trigger simplification-reviewer. </commentary> </example> <example> Context: User is reviewing a pipeline with the /review-pipeline command user: "/review-pipeline azure-pipelines.yml" assistant: "I'll use the simplification-reviewer agent to audit the pipeline for unnecessary complexity — overengineered conditions, dead steps, variables that are never used, and scripts that could be consolidated." <commentary> /review-pipeline command triggers simplification analysis as part of the comprehensive pipeline review suite. </commentary> </example>

This skill should be used when the user asks to "audit a pipeline", "review a pipeline", "review pipeline security", "set up best practices", "improve pipeline performance", "create infrastructure pipelines", "set up Terraform pipelines", "set up Bicep pipelines", or any request involving pipeline quality, security hardening, organizational standards, Infrastructure as Code deployment patterns, or Azure DevOps CLI usage. Load this skill alongside pipeline-authoring for comprehensive guidance.

This skill should be used when the user asks to "create a pipeline", "write pipeline YAML", "add a stage", "add a job", "add steps", "create a template", "debug a pipeline", "audit a pipeline", "manage variables", "configure triggers", "set up deployment", or any request involving writing, modifying, or understanding Azure DevOps pipeline YAML. Load this skill before generating or editing any Azure DevOps pipeline configuration.