AGENT SKILL

app-plan-feature-stories

Decompose a piece of work into a 3-tier Feature > Story > Task plan tree, written to disk as self-contained markdown files. Use when the user asks to plan a feature, break down work, write an implementation plan, or create a task tree for a larger effort.

License 许可
MIT
最近更新
2026-08-25
Stars
★ 0
Forks
0

01 前言 Front Matter

name
app-plan-feature-stories
tier
app
description
Decompose a piece of work into a 3-tier Feature > Story > Task plan tree, written to disk as self-contained markdown files. Use when the user asks to plan a feature, break down work, write an implementation plan, or create a task tree for a larger effort.

03 SKILL.md 正文

Plan: Feature > Stories > Tasks

Decompose work into a 3-tier hierarchy written to disk. Each file is self-contained and independently actionable.

This skill mirrors the slash command /plan-feature-story-task. Use it when a model does not auto-load slash commands or the user invokes the skill explicitly.

Input

The user provides a description of work to be done. It may be vague or detailed. Understand it deeply before decomposing.

Workflow

Phase 0: Name the Plan Directory

Format: <suggested_name>-<date>-<session_id>

  1. <suggested_name> — short kebab-case slug (max 4 words) derived from the work (e.g., auth-middleware-rewrite, api-rate-limiting)
  2. <date> — today's date in YYYYMMDD (e.g., 20260416)
  3. <session_id> — short UUID v4 prefix, 8 hex chars (e.g., a1b2c3d4)

Example: auth-middleware-rewrite-20260416-a1b2c3d4

Place under the project root at ./plans/<suggested_name>-<date>-<session_id>/. All references below to <plan-dir>/ point here. If the user specifies a different location, use that instead.

Phase 1: Understand

  1. Read the user's request carefully.
  2. Explore the codebase to understand relevant code, architecture, patterns.
  3. Ask clarifying questions if scope, goals, or constraints are ambiguous. Do not proceed until context is sufficient.

Phase 2: Decompose

Three levels:

Feature — a distinct, deliverable capability. "What does the user/system gain when this is done?"

  • Independently shippable
  • Named by capability delivered, not implementation detail

Story — a behavior slice within a feature. "What specific behavior changes when this story is complete?"

  • One observable change (user-facing or system-facing)
  • Orderable by dependency within a feature
  • Completable in a single focused session

Task — an atomic implementation step within a story. "What exact code change do I make, and how do I verify it worked?"

  • Targets specific files with specific changes
  • Minutes to hours, not days
  • Includes verification steps (test, behavior check, command)

Phase 3: Write the Plan

Create this tree under ./plans/<plan-dir>/:

<plan-dir>/
├── README.md
├── features/
│   ├── 01-<feature-name>.md
│   └── 02-<feature-name>.md
├── stories/
│   ├── 01/
│   │   ├── 01-<story-name>.md
│   │   └── 02-<story-name>.md
│   └── 02/
│       └── 01-<story-name>.md
└── tasks/
    ├── 01/
    │   ├── 01/
    │   │   ├── 01-<task-name>.md
    │   │   └── 02-<task-name>.md
    │   └── 02/
    │       └── 01-<task-name>.md
    └── 02/
        └── 01/
            └── 01-<task-name>.md

Naming: <NN>-<kebab-case-name>.md. Numeric prefix preserves ordering. Stories and tasks are grouped by their parent's feature number.

File Templates

<plan-dir>/README.md

# Plan: <Plan Title>

## Goal

<What this plan achieves — 2-3 sentences>

## Scope

<In scope and explicitly out of scope>

## Features

| #   | Feature                                     | Stories | Description      |
| --- | ------------------------------------------- | ------- | ---------------- |
| 01  | [Feature Name](features/01-feature-name.md) | N       | One-line summary |

## Execution Order

<Recommended order, noting dependencies>

<plan-dir>/features/NN-feature-name.md

# Feature: <Feature Title>

## Goal

<Capability this feature delivers — 1-2 sentences>

## Acceptance Criteria

- [ ] <Observable outcome 1>
- [ ] <Observable outcome 2>

## Stories

| #   | Story                                        | Tasks | Description      |
| --- | -------------------------------------------- | ----- | ---------------- |
| 01  | [Story Name](../stories/NN/01-story-name.md) | N     | One-line summary |

## Notes

<Architectural decisions, constraints, context specific to this feature>

<plan-dir>/stories/NN/NN-story-name.md

# Story: <Story Title>

**Feature:** [<Parent Feature Name>](../../features/NN-feature-name.md)

## Context

<Why this story exists and what behavior it changes — enough to understand without reading the parent>

## Acceptance Criteria

- [ ] <Testable behavior 1>

## Dependencies

- <Other stories this depends on, with paths, or "None">

## Tasks

| #   | Task                                           | Description      |
| --- | ---------------------------------------------- | ---------------- |
| 01  | [Task Name](../../tasks/NN/NN/01-task-name.md) | One-line summary |

<plan-dir>/tasks/NN/NN/NN-task-name.md

# Task: <Task Title>

**Story:** [<Parent Story Name>](../../../stories/NN/NN-story-name.md)

## Context

<What this task accomplishes and why — enough to execute without reading parent files>

## Target Files

- `<path/to/file.ext>` — <what changes>

## Changes

<Specific code changes. Concrete: function names, data structures, logic flow.>

## Verification

- [ ] <Command to run, test to check, behavior to observe>

Rules

  1. Self-contained files — every file includes enough context to act on independently.
  2. Relative links — all cross-references use relative markdown links.
  3. No empty shells — every field has real content. If you cannot fill a field, decomposition is not detailed enough — investigate further.
  4. Concrete tasks — name specific files and describe specific changes. "Update the config" is not a task. "Add rateLimitPool field to src/config/schema.mjs with Zod enum validation" is a task.
  5. Verify before writing — confirm file paths, function names, and patterns by reading actual code before referencing them.

After Writing

Summarize to the user:

  • Total count: N features, N stories, N tasks
  • Recommended execution order
  • Risks, open questions, or decisions needing user input

Reference

Source command: .claude/commands/plan-feature-story-task.md.

04 同仓库更多 Skills

carlosmarte
carlosmarte/ai-agent-artifacts

validation_handoff

Forces the agent to conclude its response by generating a strict validation checklist and outlining pending actions, focusing heavily on defensive execution and edge-case testing.

★ 0
carlosmarte
carlosmarte/ai-agent-artifacts

project-ployglot-project-scafford

Scaffold a polyglot library project as TypeScript + Python twin packages under `packages/ts/` + `packages/py/`, with a shared `SPEC.md`, cross-language `tests/parity/fixtures.json`, side-by-side `examples/sdk/` + `examples/api/` documentation, per-package `Makefile`, and a root orchestrator that exposes `make ci`. Open-source friendly — no organization-specific paths or scopes baked in. Use when starting a new polyglot library, porting a single-language library to a twin, or auditing/regularizing an existing two-language repo against the canonical layout.

★ 0
carlosmarte
carlosmarte/ai-agent-artifacts

playwright-ui-locator

Auto-heal Playwright tests after a frontend developer changes a div's class, id, or other selector. Catches a failing Playwright test, captures the old vs new DOM tree, finds the element that drifted, derives the most stable replacement locator, and proposes the fix as a GitHub pull request (confident heals) or files a GitHub issue (ambiguous / element-removed cases) — never editing the working tree directly. Trigger when Playwright tests fail with "locator resolved to 0 elements", "Timeout waiting for selector", or after a UI refactor / class rename / Tailwind cleanup broke previously-passing tests.

设计师代码审查测试与质量
★ 0
carlosmarte
carlosmarte/ai-agent-artifacts

org-learn-context-window

Distill the current conversation into a reusable command or skill artifact. Use when the user wants to capture a workflow they just executed (create a reusable slash command) or formalize a resolution/mistake/correction into a reusable skill so the same class of issue never recurs.

Skill与智能体开发
★ 0
carlosmarte
carlosmarte/ai-agent-artifacts

org-agent-skill-creator

Scaffold and validate scope-aware agent skills (company, enterprise, org, team, project, application). Use when the user asks to create a new agent skill, generate a SKILL.md file, scaffold a skill directory, or validate that existing skill directories meet SKILL_SPEC.md (frontmatter, naming rules, structure, scope/tier).

Skill与智能体开发
★ 0
carlosmarte
carlosmarte/ai-agent-artifacts

app-workshop-session-pdf-recap-next

Extract content from a workshop session PDF and produce a structured recap focused on (1) what was learned, (2) next steps, and (3) possible enhancements. Ingests a single PDF file path, parses session material (slides, transcripts, exercises, Q&A), and emits a Markdown report with Learning Recap, Next Steps, and Enhancement Opportunities sections with page-level traceability. Use when the user asks to "summarize a workshop", "recap a training PDF", "pull next steps from a session deck", or provides a workshop/seminar/tutorial PDF and wants a learn-and-act summary.

文档处理
★ 0