What problem does it solve? Teams building LLM-powered systems often waste weeks automating tasks the model cannot do, over-engineering tool scaffolding, or discovering token cost overruns only in production. This Skill provides a decision methodology for validating task-model fit, structuring multi-stage pipelines, and estimating costs before committing to implementation. ## Core Features & Use Cases - Task-Model Fit Validation: Proceed/stop criteria tables plus a manual prototype step to verify the model can handle the task before any automation is built. - Pipeline Architecture Patterns: A canonical acquire → prepare → process → parse → render structure with file-system-based state management, idempotent stages, and a ready-to-customize Python pipeline template. - Cost and Scale Estimation: Token-counting formulas with a 20-30% buffer, plus guidance on single-agent vs multi-agent selection and architectural reduction. - Use Case: Before building a batch job that grades 900 documents with an LLM, use this Skill to validate the task with one manual example, scaffold the staged pipeline from the included template, and project the total dollar cost. ## Quick Start Ask the agent to evaluate whether your planned LLM task is a good model fit and to sketch a staged pipeline with a cost estimate using the project-development methodology.
