Skip to content
OpenAgentSkill
Browsing is open · Catalog captured 2026-11-21 · Automatic updates and account services are paused.
Browser & automation

create-tasks

Generate implementation tasks from architecture document. Usage /groundwork:create-tasks

etr · Static checks on file

Installation

Read the source and dependencies before running a command. Copying does not install or execute anything.

npx skills add etr/groundwork --skill create-tasks

About this skill


name: create-tasks description: Generate implementation tasks from architecture document. Usage /groundwork:create-tasks argument-hint: "[filter]" allowed-tools: ["Read", "Edit", "Write", "Glob", "Grep", "Task", "Skill"]

Task Generation Skill

Translates product specs and architecture into actionable implementation tasks.

Pre-flight: Model Recommendation

Your current effort level is {{effort_level}}.

Skip this step silently if effort is high, xhigh, or max (the scale is low < medium < high < xhigh < max, so xhigh and max are already above high) AND you are Opus (1M context). If effort is low or medium (i.e. below high), you MUST show the recommendation prompt — regardless of model. If you are not Opus (1M context), you MUST show the recommendation prompt - regardless of effort level.

Otherwise → use AskUserQuestion:

{
  "questions": [{
    "question": "Do you want to switch? Task granularity judgment and vertical slicing decisions determine execution success.\n\nTo switch: cancel, run `/model opus[1m]` and `/effort high`, then re-invoke this skill.",
    "header": "Recommended: Opus (1M context) at high effort",
    "options": [
      { "label": "Continue" },
      { "label": "Cancel — I'll switch first" }
    ],
    "multiSelect": false
  }]
}

If the user selects "Cancel — I'll switch first": output the switching commands above and stop. Do not proceed with the skill.

Step 0: Resolve Project Context

Before loading specs, ensure project context is resolved:

  1. Monorepo check: Does .groundwork.yml exist at the repo root?
    • If yes → Is {{project_name}} non-empty?
      • If empty → Invoke Skill(skill="groundwork:select-project") to select a project, then restart this skill.
      • If set → Project is {{project_name}}, specs at {{specs_dir}}/.
    • If no → Continue (single-project repo).
  2. CWD mismatch check (monorepo only):
    • Skip if not in monorepo mode or if the project was just selected in item 1 above.
    • If CWD is the repo root → fine, proceed.
    • Check which project's path CWD falls inside (compare against all projects in .groundwork.yml).
    • If CWD is inside the selected project's path → fine, proceed.
    • If CWD is inside a different project's path → warn via AskUserQuestion:

      "You're working from <cwd> (inside [cwd-project]), but the selected Groundwork project is [selected-project] ([selected-project-path]/). What would you like to do?"

      • "Switch to [cwd-project]"
      • "Stay with [selected-project]" If the user switches, invoke Skill(skill="groundwork:select-project").
    • If CWD doesn't match any project → proceed without warning (shared directory).
  3. Proceed with the resolved project context. All {{specs_dir}}/ paths will resolve to the correct location.

File Locations

  • Input:
    • {{specs_dir}}/product_specs.md (PRD with EARS requirements)
    • {{specs_dir}}/architecture.md (architecture with decision records)
    • {{specs_dir}}/design_system.md (design system with DP/BRD/UXD decisions) - optional
  • Output: {{specs_dir}}/tasks/ directory with per-milestone subdirectories:
    specs/tasks/
    ├── _index.md          # Overview, milestone summary, dependency graph
    ├── M1-core-auth/
    │   ├── TASK-001.md
    │   ├── TASK-002.md
    │   └── TASK-003.md
    ├── M2-upload/
    │   ├── TASK-004.md
    │   └── TASK-005.md
    └── parking-lot.md     # Deferred tasks
    

Workflow Overview

  1. Load Context - Read PRD and architecture
  2. Identify Milestones - Define testable product milestones
  3. Generate Tasks - Break down into implementable tasks
  4. Map Dependencies - Establish task ordering
  5. Review & Refine - Iterate with user on task breakdown

Important: When asking questions to the user (milestone validation, task review, scope clarification), always use the AskUserQuestion tool with appropriate options.

Step 1: Load Context

Read the input specs (each may be a single file or directory) and extract:

From PRD:

  • Single file: {{specs_dir}}/product_specs.md
  • Directory: {{specs_dir}}/product_specs/ (aggregate all .md files)
  • Extract: Feature list with EARS requirements, Non-functional requirements, Release strategy (Alpha → Beta → GA)

From Architecture:

  • Single file: {{specs_dir}}/architecture.md
  • Directory: {{specs_dir}}/architecture/ (aggregate all .md files)
  • Extract: Component list and responsibilities, Technology choices, Decision records (understand constraints)

From Design System (if exists):

  • Single file: {{specs_dir}}/design_system.md
  • Extract: Design principles (DP-NNN), Brand decisions (BRD-NNN), UX patterns (UXD-NNN)
  • Use for: Design-related tasks, component styling tasks, accessibility requirements

Detection: Check for file first (takes precedence), then directory. When reading a directory, aggregate all .md files recursively with _index.md first, then numerically-prefixed files, then alphabetically.

If PRD or architecture is missing, prompt user:

"I need both the PRD and architecture to generate tasks.

  • PRD missing? Run /groundwork:design-product
  • Architecture missing? Run /groundwork:design-architecture"

If design system exists, incorporate design context into UI/frontend tasks.

Step 2: Check Existing Tasks and Milestones

Before defining milestones, check if tasks already exist:

  1. Check for {{specs_dir}}/tasks/_index.md or {{specs_dir}}/tasks.md
  2. If found, read to understand existing milestones, task numbering, and status

If existing tasks found:

  • List existing milestones and their status (complete, in progress, not started)
  • Determine the highest task number (e.g., TASK-047) so new tasks continue the sequence
  • Determine the highest milestone number (e.g., M5) so new milestones continue the sequence
  • Present to user: "Found existing milestones M1-M5 with 47 tasks. New tasks will be added as TA
Read the complete source ↗