task-planner

Guidelines for transforming descriptive tasks into technical breakdowns with main tasks (product-focused, well-formatted JIRA tickets) and subtasks (implementation-focused, technical chunks). Scope: task generation, consolidation, FE/BE separation, point estimation (0.1-5 scale), acceptance criteria. Triggers: "task", "breakdown", "subtask", "main task", "ticket", "JIRA", "estimation", "point".

$ Installer

git clone https://github.com/warpcode/dotfiles /tmp/dotfiles && cp -r /tmp/dotfiles/generic/.config/opencode/skill/task-planner ~/.claude/skills/dotfiles

// tip: Run this command in your terminal to install the skill


name: task-planner description: >- Guidelines for transforming descriptive tasks into technical breakdowns with main tasks (product-focused, well-formatted JIRA tickets) and subtasks (implementation-focused, technical chunks). Scope: task generation, consolidation, FE/BE separation, point estimation (0.1-5 scale), acceptance criteria. Triggers: "task", "breakdown", "subtask", "main task", "ticket", "JIRA", "estimation", "point".

Task Planner

METHOD

Phase 1: Clarification

Detect task type from user input:

  • "Create task", "task", or "main task" → Main Task Mode
  • "Breakdown", "subtask", or "breakdown of task" → Subtask Mode
  • Default: "task" alone → Main Task Mode
  • IF description missing OR ambiguous → Ask for clarification

Phase 2: Planning

IF Main Task Mode → READ FILE: @references/main-task-structure.md → Generate ticket structure → Generate acceptance criteria → Ask user for format choice. IF Subtask Mode → READ FILE: @references/subtask-guidelines.md → Generate subtasks → READ FILE: @references/estimation-rubric.md → Estimate points → READ FILE: @references/consolidation-rules.md → Validate.

Phase 3: Execution

Main Task Workflow

  1. Detect ticket type (User Story / Technical Ticket / Bug Report).
  2. Generate well-formatted ticket structure.
  3. Generate Gherkin scenarios (if applicable).
  4. Ask user: Acceptance criteria format (Gherkin OR Checklist OR Custom).
  5. Ask user: Resources section (auto-fill codebase search OR manual).
  6. Output final main task.

Subtask Workflow

  1. Generate subtasks with title and description (technical, granular).
  2. Estimate points (0.1-5 scale: Complexity + Effort + Uncertainty).
  3. Validate consolidation/breakdown (target ≤3 points).
  4. READ FILE: @references/fe-be-separation.md → Separate FE/BE.
  5. Output subtask list (format: Title [X points] - Description).

Phase 4: Validation

Verify: Main tasks contain required sections (Title, Overview, Description, Acceptance Criteria, Technical Brief, Resources). Verify: Subtasks ≤3 points (rarely 5). Verify: FE/BE separation applied. Verify: Point estimates follow rubric.

EXAMPLES

Example 1: Create main task

User: "Create main task for daily data export cron job" Agent: Generates well-formatted JIRA ticket with:

  • Title: "Implement cron script for daily data export"
  • Overview: User story format
  • Description: Requirements, scope, constraints
  • Technical Brief: Executive Summary, Impacted Areas, Risks
  • Acceptance Criteria: Gherkin scenarios (pending user confirmation)
  • Resources: Ask user (auto-fill OR manual)

Example 2: Break down task

User: "Break down cron export task into subtasks" Agent: Generates subtasks:

Backend

  • Add export jobs table [1 point]
    • Add export_jobs table with columns: id, user_id, status, file_path, created_at
  • Create export trigger endpoint [1 point]
    • Implement POST /api/exports/trigger endpoint with user permission validation

Frontend

  • Create export status page [1 point]
    • Create page component with export history table and pagination
  • Add export trigger button [0.5 points]
    • Add button to manually trigger export with loading state

CONSTRAINTS

  1. Main tasks: Well-formatted JIRA tickets with required sections.
  2. Subtasks: Technical, granular, ≤3 points, with title and description for every subtask.
  3. Point scale: 0.1, 0.5, 1, 2, 3, 4, 5 (1 point = 1 day).
  4. FE/BE: Always separate frontend and backend subtasks (NEVER share).
  5. Consolidation: Merge related work to avoid excessive granularity.
  6. Validation: Check subtask points → consolidate IF >3 OR split IF <0.1.