from turkey-build71
Multi-agent app builder and orchestrator: bootstraps projects, creates feature branches, and runs runtime-verified builds with a PM conductor and QA gates.
Turkey Build is an orchestrator skill that coordinates multiple specialized agents to produce production-ready applications. It bootstraps a repository, manages feature/bugfix branches, assigns agent roles (PM, backend, frontend, QA, Visual QA), and enforces a high-quality conductor gate (runtime verification and visual QA) before releases. The workflow covers greenfield builds, iterative feature work, bugfixes, refactors, migrations, UI polish, and audits.
Invoke when you need to generate or modify an application end-to-end ("build me X"). Use for new projects, scoped feature work, migration plans, or automated bugfix flows where runtime verification and visual QA are required. It's intended for environments where agents have filesystem and CI access.
references/ docs but does not bundle runnable scripts in SKILL.mdreferences/ directory (ORCHESTRATION, PM-AGENT, E2E, Visual QA, etc.) required to run safelyAgents that can run Bash, read/write files, and orchestrate git/CI (Claude Code, workspace agents with shell access, similar multi-agent frameworks).
Multi-agent app builder/orchestrator with 7 modes (greenfield, iteration, bugfix, refactor, UI polish, migration, audit). PM agent coordinates specialized agents via feature branches with 98% quality gate and runtime verification. No bundled scripts — all logic lives in 17+ reference files. Requires Claude Code + aimem MCP server.
Well-structured orchestration skill. Modifies .claude/settings.local.json during bootstrap — standard for this type of tool but worth noting. No scripts to execute or test. Heavy reliance on external reference files means the skill's value depends entirely on those files being present and well-written.
