ai-agent-workflowlisted

# AI Agent Workflow Use this skill when the goal is not just to get one answer, but to build a repeatable AI-assisted workflow and the main stack direction is already mostly known. This skill is for taking AI ideas and turning them into structured loops: prompts, tools, retrieval, checks, and reusable building blocks. ## Inputs Useful inputs for this skill include: - the repeatable task you want the workflow to handle - current prompt, script, tool, or skill draft if one exists - model constraints such as local-only or API limits when the basic stack is already chosen - available tools, MCP servers, files, or data sources - quality bar, failure modes, and evaluation expectations ## Outputs Strong outputs from this skill usually include one or more of: - a reusable workflow specification - a prompt template or prompt stack - a tool integration or MCP plan - an evaluation loop or quality rubric - a recommendation for whether this should be a prompt, skill, script, or MCP server - a minimal implementation path with clear next steps ## Non-goals This skill is not the best fit for: - choosing between local deployment stacks such as MLX, GGUF, LM Studio, Ollama, or vLLM - hardware-first local LLM decisions or local serving setup questions - creating a new skill from scratch or rewriting the skill artifact itself - benchmarking a skill, improving skill triggering, or building a skill eval harness - one-off content writing with no reusable workflow need - generic code fixes u