Refine draft task specifications into implementation-ready tasks via multi-agent workflow.
npx skills add https://github.com/Gamezar/opencode-cek --skill sdd-plan-gamezar
Please help me install this Agent Skill. Skill: sdd-plan Source: https://github.com/Gamezar/opencode-cek/tree/main/plugins/sdd/skills/sdd-plan Command: npx skills add https://github.com/Gamezar/opencode-cek --skill sdd-plan-gamezar
This Skill automates the complex process of refining draft task specifications into actionable, well-planned implementation steps, ensuring quality and efficiency through a multi-agent workflow.
Refine the draft task specification located at '.specs/tasks/draft/new-feature.feature.md' using the sdd-plan skill.
High-intent search queries and answers about installing and using this skill.
To refine draft task specifications, you can use a multi-agent workflow that conducts parallel analysis, synthesizes architecture, breaks down tasks, and verifies each phase with LLM-as-Judge quality gates to produce actionable plans.
The best way to parallelize tasks from a vague draft is by running concurrent research, codebase analysis, and business analysis, then synthesizing the results into an architecture that can be decomposed into parallelizable implementation steps.
An LLM-as-Judge quality gate validates output at each stage of the task decomposition workflow, ensuring draft specifications meet high technical standards before being parallelized and verified as implementation-ready steps.
Yes, you can use human-in-the-loop checkpoints during task specification refinement to manually verify phases, and the workflow supports incremental refinement via git diff to track specification changes.
Yes, multi-agent workflow automation supports incremental refinement using git diff, allowing you to track and process only the modified parts of draft task specifications through the verification phases.
You need specification decomposition when transforming vague product feature drafts into detailed, actionable, and parallelizable implementation steps that require integrated research, codebase analysis, and business analysis.
TrommlOrchestrate a six-stage planning pipeline to transform feature ideas into validated task specifications.
anexpnGenerate and update problem_statement.md, decisions.md, runbook.md, and progress.md for iterative development.
466852675Evaluate AI agent performance using multi-dimensional rubrics and LLM-as-judge.
acarmonagOrchestrate multi-agent software development across specification, architecture, and implementation phases.
