executing-plans
Execute a plan step-by-step with checkpoints and verification at each stage
Execute a plan step-by-step with checkpoints and verification at each stage
npx skills add https://github.com/obra/superpowers/tree/main/skills/executing-plansExecute a plan step-by-step with checkpoints and verification at each stage
This repo contains 13 individual skills — each has its own dedicated page.
Structured ideation and problem decomposition frameworks
Split work across parallel subagents and coordinate their outputs
Branch close checklist: tests, commit message, pull request, and review request
Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation
Prepare code for review: self-review, test coverage, and pull request description
Orchestrate specialized subagents for different parts of a task
Hypothesis-driven debugging loop: observe, hypothesize, test, verify
TDD loop: write the failing test first, implement the minimal change, verify, then refactor
Use git worktrees to run parallel agent sessions on separate branches
Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions
Force a verification pass before any task is marked done
Write structured implementation plans before starting complex tasks
Use when creating new skills, editing existing skills, or verifying skills work before deployment
Claude Code skill: execute implementation plans with fresh subagents per task, two-stage review gates, and parallel mode via worktree isolation. Based on obra/superpowers, parallel mode adapted from theradengai/spawn-tasks.
一个可复用的 Codex skill，用来把长任务按以下流程运行： - `executive orchestrator`：控制全流程 - `continuation planner`：先接管半成品项目现状 - `phase planner`：按阶段规划 - `task planner`：按任务切片规划 - `implementer`：实现当前任务 - `validator`：验证结果 - `clarification manager`：把非阻塞疑问先记录，当前切片完成后再集中补问
An agentic project manager for AI-assisted coding — persistent memory, living roadmaps, and automatic architecture tracking
Execution-Grounded Task-State Runtime — obligation tracking, stale quarantine, and resume safety plugin for Claude Code
🤖 AI Agent-driven Kaggle competition workflow. Battle-tested patterns for score stabilization, submission troubleshooting, kernel workflows, and spec-driven development.
Language-neutral multi-model (Claude Opus + GPT/Codex) workflow scaffold: 7 wf-* user-invocable skills + 3-lane (shell/python/node) governance adapters + zero-dep ./tasks.sh dispatcher.