Work with Obsidian vaults (plain Markdown notes) and automate via obsidian-cli.
npx skills add Hisham-Hussein/claude-forge --skill "create-execution-plan"
Install specific skill from multi-skill repository
# Description
Use when you have architecture documents and user stories ready and need to plan implementation phases. Invoke after design docs are complete, before coding begins, or when asked to create an execution plan, project roadmap, or phase breakdown.
# SKILL.md
name: create-execution-plan
description: Use when you have architecture documents and user stories ready and need to plan implementation phases. Invoke after design docs are complete, before coding begins, or when asked to create an execution plan, project roadmap, or phase breakdown.
This skill produces project-level roadmaps, NOT implementation-level plans.
| Level | Skill | Output |
|---|---|---|
| Project (this skill) | /create-execution-plan | Phases → Stories → Tasks (what to build) |
| Implementation | superpowers:writing-plans | Task → Steps (how to build each task) |
| Execution | superpowers:executing-plans | Steps → Code |
Methodology: Clean Architecture + Vertical Slices + FDD
- Clean Architecture layers: Domain → Application → Infrastructure (dependency flows inward)
- Vertical Slices: Each phase delivers working end-to-end functionality
- FDD-style tasks: Atomic tasks with explicit Input/Output/Test
Task Structure (CRITICAL):
Every task MUST have:
Task X.Y: [Description]
├── Input: What this task needs to start
├── Output: What this task produces
├── Test: How to verify correctness
└── Commit: Conventional commit message
This structure minimizes AI coding errors by making requirements explicit.
Transform architecture documents and user stories into phased execution plans with atomic, TDD-ready tasks. Output enables AI-assisted development with clear work units that fit within context windows and have verifiable success criteria.
Required inputs:
- USER-STORIES.md (with acceptance criteria)
- ARCHITECTURE-DOC.md or Design Doc (from /create-design-doc)
Optional inputs:
- BUSINESS-CASE.md (constraints, success criteria)
- Tech stack research, API docs, data templates
Output:
- EXECUTION-PLAN.md (phases, stories, tasks with I/O/Test)
- jira-import.csv (optional JIRA bulk import)
Run: /create-execution-plan path/to/USER-STORIES.md path/to/ARCHITECTURE-DOC.md
Confirm inputs:
- Where are the required documents?
- USER-STORIES.md: [path]
-
ARCHITECTURE-DOC.md: [path]
-
Any optional documents?
- BUSINESS-CASE.md: [path or skip]
- Tech stack research: [path or skip]
-
Other (API docs, data templates): [paths or skip]
-
Where should output go?
- Default: same directory as USER-STORIES.md
-
Custom: [path]
-
Generate JIRA export?
- Yes / No
After confirming, read workflows/generate-plan.md and follow it exactly.
| Intent | Workflow |
|---|---|
| Generate new plan | workflows/generate-plan.md |
| Direct invocation with args | workflows/generate-plan.md (parse paths from args) |
If arguments provided: Parse document paths from ARGUMENTS and proceed directly to workflow.
All in references/:
- Methodology: methodology.md (Clean Architecture + VSA + FDD integration)
- Task decomposition: task-decomposition.md (FDD-style atomic tasks)
All in templates/:
- Plan structure: execution-plan.md (EXECUTION-PLAN.md template)
- JIRA export: jira-import.csv (CSV format for bulk import)
Execution plan is complete when:
- [ ] All user stories assigned to phases
- [ ] Must-have stories in early phases
- [ ] Each phase delivers working vertical slice
- [ ] Every task has Input/Output/Test
- [ ] Phase transitions documented (output → input)
- [ ] Definition of Done per phase
- [ ] JIRA export generated (if requested)
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.