Generate task-level implementation plans with exact file paths and code samples.
npx skills add https://github.com/samjhecht/wrangler --skill writing-plans-samjhecht
Please help me install this Agent Skill. Skill: writing-plans Source: https://github.com/samjhecht/wrangler/tree/main/skills/writing-plans Command: npx skills add https://github.com/samjhecht/wrangler --skill writing-plans-samjhecht
Vague or incomplete implementation plans lead to confusion, rework, and delays, especially when executed by AI agents or engineers unfamiliar with the codebase.
The design for the 'user-onboarding' feature is approved. Create a detailed implementation plan for it.
High-intent search queries and answers about installing and using this skill.
Generate a comprehensive implementation plan by breaking the approved design into small, actionable tasks (2-5 minutes each) with exact file paths, complete code examples, precise commands with expected outputs, and acceptance criteria. This ensures engineers or AI agents can execute without ambiguity or rework.
A complete plan includes exact file paths, full code samples, testing steps, verification commands with expected outputs, five TDD steps, acceptance criteria, and cross-task references. This eliminates guesswork for executors unfamiliar with the codebase.
Use task-level plans when a design is complete and ready for execution by engineers or AI agents who have minimal codebase context. They prevent confusion, rework, and delays by documenting everything needed for accurate implementation.
Yes. Plans designed with zero-context assumptions—exact file paths, complete code examples, precise commands, and expected outputs—enable AI agents to execute accurately without domain knowledge or codebase familiarity.
A design describes what to build and why; an implementation plan details how to build it step by step. Plans add executable specifics: file paths, code, commands, outputs, and MCP-tracked tasks that a design alone doesn't provide.
Break the implementation into bite-sized tasks by isolating one logical unit per task: a single file modification, a test addition, or one command execution. Include all context—exact paths, complete code, expected results—so each task is self-contained and fast to execute.
SergeiGolosGuide architectural decisions through requirements analysis, trade-off evaluation, and ADR documentation.
duronextEnforce verification command execution and evidence review before completion claims.
awood45Implement planned tasks with strict test-driven development and human approval.
a-ariffCreate detailed implementation plans with granular tasks and exact file paths.
