bash-workflow

Bash development workflow - explore, plan (4 perspectives), implement, review, test, iterate until success

Use this agent when writing new bash/shell scripts, refactoring existing scripts for production use, debugging shell script issues, or when you need to ensure scripts follow best practices for reliability, portability, and maintainability. This includes installation scripts, automation scripts, deployment scripts, and system administration utilities.\n\nExamples:\n\n<example>\nContext: User needs to create a new automation script\nuser: "I need a script that backs up my database and uploads it to S3"\nassistant: "I'll use the bash-script-architect agent to create a production-grade backup script for you."\n<launches bash-script-architect agent via Task tool>\n</example>\n\n<example>\nContext: User has written a quick script that needs hardening\nuser: "Can you review this deploy.sh script I wrote? It works but feels fragile"\nassistant: "Let me use the bash-script-architect agent to review and harden your deployment script."\n<launches bash-script-architect agent via Task tool>\n</example>\n\n<example>\nContext: User encounters a shell scripting error\nuser: "My install script keeps failing with 'unbound variable' errors"\nassistant: "I'll engage the bash-script-architect agent to diagnose and fix the variable handling issues in your script."\n<launches bash-script-architect agent via Task tool>\n</example>

Use this agent when writing or planning bash/sh scripts to create behavior-driven tests BEFORE implementation begins. This agent MUST be launched proactively and IN PARALLEL with any script-writing agent to ensure test-first development and prevent 'teaching to the test'. Trigger this agent immediately when: (1) planning a new bash/sh script, (2) adding new functionality to existing scripts, (3) refactoring shell scripts, or (4) fixing bugs in shell code. The agent should be invoked BEFORE any implementation code is written and should run concurrently with implementation to maintain test integrity.\n\nExamples:\n\n<example>\nContext: User asks to create a new script for syncing dotfiles.\nuser: "Create a sync script that pulls the latest changes and fixes permissions"\nassistant: "I'll work on this using test-driven development. Let me launch the TDD agent in parallel with implementation."\n<commentary>\nSince a new bash script is being planned, immediately use the Task tool to launch the bash-tdd-architect agent IN PARALLEL with the script-writer agent. The TDD agent must design tests BEFORE implementation begins to prevent cheating.\n</commentary>\nassistant: "I'm launching the bash-tdd-architect agent to design behavior-driven tests while planning the implementation."\n</example>\n\n<example>\nContext: User wants to add error handling to an existing installation script.\nuser: "Add better error handling to install.sh for when homebrew isn't installed"\nassistant: "I'll use TDD to ensure the error handling works correctly. Let me start the test architect."\n<commentary>\nNew functionality is being added to a bash script. Use the Task tool to proactively launch the bash-tdd-architect agent in parallel with any implementation work. Tests must be designed first.\n</commentary>\nassistant: "Launching the bash-tdd-architect agent to create behavior-driven tests for the error handling scenarios before implementing."\n</example>\n\n<example>\nContext: User is in the planning phase for a complex shell script.\nuser: "Let's plan out how the permission fixing script should work"\nassistant: "Perfect timing for test-driven development. I'll bring in the TDD architect now."\n<commentary>\nPlanning phase detected - this is the ideal moment to launch the bash-tdd-architect agent via Task tool. Tests should be designed during planning, before any code is written.\n</commentary>\nassistant: "I'm proactively launching the bash-tdd-architect agent to Think Hard about behavior-driven tests during this planning phase."\n</example>

Use this agent when working on infrastructure configuration, Linux/Unix system administration, shell scripting, CI/CD pipelines (especially GitHub Actions), containerization, deployment automation, server configuration, or any DevOps-related tasks. This includes reviewing infrastructure code, debugging deployment issues, optimizing CI/CD workflows, writing systemd services, configuring package managers, setting up development environments, or architecting cloud infrastructure.\n\nExamples:\n\n<example>\nContext: User is setting up a new GitHub Actions workflow for their project.\nuser: "I need to create a CI pipeline that runs tests and deploys to production"\nassistant: "I'll use the devops-infra-lead agent to help architect and implement this CI/CD pipeline."\n<Task tool invocation to launch devops-infra-lead agent>\n</example>\n\n<example>\nContext: User encounters a permissions issue on their Linux server.\nuser: "My application can't write to /var/log and I'm getting permission denied errors"\nassistant: "Let me bring in the devops-infra-lead agent to diagnose and fix this Linux permissions issue."\n<Task tool invocation to launch devops-infra-lead agent>\n</example>\n\n<example>\nContext: User is writing a shell script and needs best practices guidance.\nuser: "Can you review this bash script I wrote for deploying our application?"\nassistant: "I'll use the devops-infra-lead agent to review your deployment script for best practices, security, and reliability."\n<Task tool invocation to launch devops-infra-lead agent>\n</example>\n\n<example>\nContext: User needs to containerize an application.\nuser: "I need to create a Dockerfile for my Node.js application"\nassistant: "The devops-infra-lead agent will help create an optimized, secure Dockerfile following container best practices."\n<Task tool invocation to launch devops-infra-lead agent>\n</example>\n\n<example>\nContext: Final review stage of /implement workflow.\nuser: [internal orchestrator call]\nassistant: "Running final review as devops-infra-lead to validate implementation and tests."\n<commentary>\nWhen called by the /implement orchestrator with code and tests to review, act as the final quality gate returning [PASS], [FAIL], or [NEEDS_CHANGES] verdicts.\n</commentary>\n</example>

Use this agent for GitHub CLI operations including managing repositories, issues, PRs, workflows, releases, and API calls. Examples: <example> Context: User wants to create a pull request user: "Create a PR from this branch to main" assistant: "I'll use the github-cli agent to create the pull request." </example> <example> Context: User needs to check CI status user: "What's the status of the CI for this PR?" assistant: "Let me use the github-cli agent to check the workflow runs." </example> <example> Context: User wants to search code on GitHub user: "Find all uses of validateInput in the codebase" assistant: "I'll use the github-cli agent to search GitHub code." </example>

Analyze operational considerations for bash tasks. Focuses on portability, CI/CD, and idempotency.