Two-phase workflow for technical specifications - WRITING phase creates and commits specs then stops; IMPLEMENTATION phase follows specs, updates status, and commits all changes. Different agents handle each phase.
Validate production batch execution - trigger daily runs and analyze traces for architecture completeness and result quality
Understanding the threat model for self-hosted GitHub Actions runners. GitHub-hosted vs self-hosted comparison and secure deployment patterns.
Use when distributing multiple tasks across agents or VMs - handles work manifests, dependency graphs, parallel execution, agent coordination, and failure recovery for batch development workflows
Use when you have a spec or requirements for a multi-step task, before touching code
Comprehensive GitHub code review with AI-powered swarm coordination
Checks if critical environment variables are set for the project.
Deploys a component (frontend, backend, or landing) using SST. Use when deploying to AWS, pushing to dev or prod environments.
Use when executing implementation plans with independent tasks in the current session or facing 3+ independent issues that can be investigated without shared state or dependencies - dispatches fresh subagent for each task with code review between tasks, enabling fast iteration with quality gates
Standard Operating Procedure for /optimize phase. Covers performance benchmarking, accessibility audit, security review, and code quality checks.
Validates Implementation Phase completion before advancing to Testing Phase. Checks code compilation, test coverage, implementation completeness, and code quality standards.
Creates pull requests, deployment checklists, and manages release processes
バージョン整合性チェック→提案→承認→GitHub Release作成までの標準リリースフロー（マルチエコシステム対応）
Initial Convex workspace setup in Coder workspaces with self-hosted Convex deployment, authentication configuration, Docker setup, and environment variable generation
Comprehensive management of MCP Gateway (gh-aw-mcpg) for running GitHub MCP servers. Use when: (1) Starting/stopping MCP Gateway containers with GitHub MCP server support, (2) Debugging gateway connection or authentication issues, (3) Configuring Copilot CLI to use MCP Gateway via HTTP, (4) Setting up gateway for use with AWF (Agentic Workflow Firewall), (5) Troubleshooting MCP server initialization or token passthrough problems, (6) Managing Docker-based MCP server lifecycle. Includes scripts for gateway startup, health checking, debugging, and Copilot CLI integration.
Scaffolds a new Google Antigravity project with the standard .agent harness, rules, and workflows.
Linux VM bootstrap verification skill. MUST BE USED when setting up new VMs or verifying environment. Supports modes: check (warn-only), install (operator-confirmed), strict (CI-ready). Enforces Linux-only + mise as canonical; honors preference brew→npm (with apt fallback). Verifies required tools: mise, node, pnpm, python, poetry, gh, railway, bd, tmux, jq, rg. Handles optional tools as warnings: tailscale, playwright, docker, bv. Never prints/seeds secrets; never stores tokens in repo/YAML; Railway vars only for app runtime env. Safe on dirty repos (refuses and points to dirty-repo-bootstrap skill, or snapshots WIP branch). Keywords: vm, bootstrap, setup, mise, toolchain, linux, environment, provision, verify, new vm
