Skip to content

simonepri/ifttt-lint / .agents/skills/plan

Structured design intake for non-trivial tasks. Asks clarifying questions and produces a design proposal.

Plan

Structured intake for complex tasks. The plan lives in a file, not in conversation output. Stay at the architecture level — define boundaries, interfaces, and data flow. Leave implementation details to the implementer.

  1. Understand: ask clarifying questions using AskUserQuestion — this is mandatory, never skip it. Use batches of ~4 questions, multiple-choice with 2-4 options + "Other / Not sure" when possible. Iterate until the reasoning framework is filled in — Problem, Goal, Background, Constraints. Do not write code or produce a plan until the user has answered at least one round of questions.
  2. Propose: write the proposal to a temp file (/tmp/plan-<timestamp>.md). Print a markdown link to the file (e.g., [plan-name.md](/tmp/plan-name.md)) so the user can click to open it — do not dump the plan into conversation output. Use the reasoning framework from workflow.md, structured top-down:
    • Problem(s): what's wrong or missing — the motivation for this work
    • Goal(s): what success looks like (optional if obvious from the problem)
    • Background: relevant context to understand the problem/solution — existing architecture, constraints, prior art. Use diagrams (ASCII or Mermaid) for data flow and component relationships when helpful.
    • Solution: what we're doing about it and how it integrates with what already exists. Structure as versioned milestones when the work has natural increments (v0 = minimal viable, v1 = full feature, etc.). Each version should be independently shippable. Within a version, group tasks by dependency — independent tasks are parallelizable. Define interfaces (API contracts, data shapes, protocols) at boundaries where parallel work can happen. Use diagrams (ASCII or Mermaid) for architecture and data flow when helpful.
    • Alternative(s) considered: other approaches evaluated and why not (optional, include when non-obvious)
    • Implementation details: specific file paths, configs, schemas, code snippets that clarify the solution. Optional — the plan should be understandable without it.
  3. Iterate: when the user requests changes, update the file in place. Never reprint the full plan — just summarize what changed.
  4. Confirm: wait for user confirmation before implementing.
brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).Stars: 319kUpdated Claude CodeCodexVerifiedInstall
brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).Stars: 196kUpdated Claude CodeCodexVerifiedInstall