Skip to content
cccskills
FreePublished on GitHub

task-planning

Generates phased, dependency-ordered implementation tasks from specifications. Use after spec is complete and before starting implementation.

The English summary is being prepared. Showing the original description.

See how to install

Files included (4)

  • SKILL.md4.1 KB
  • modules/dependency-patterns.md6.0 KB
  • modules/phase-structure.md4.9 KB
  • modules/tech-stack-patterns.md2.3 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Task Planning

Overview

Transforms specifications and implementation plans into concrete, dependency-ordered tasks. Creates phased breakdowns that guide systematic implementation.

When To Use

  • Converting specifications to implementation tasks
  • Planning feature implementation order
  • Identifying parallel execution opportunities
  • Breaking down complex features into phases

When NOT To Use

  • Writing specifications - use spec-writing

Task Phases

Tasks follow a 5-phase structure from setup through polish:

  • Phase 0: Setup - Project initialization, dependencies, configuration
  • Phase 1: Foundation - Data models, interfaces, test infrastructure
  • Phase 2: Core Implementation - Business logic, APIs, services
  • Phase 3: Integration - External services, middleware, logging
  • Phase 4: Polish - Optimization, documentation, final testing

For detailed phase definitions, selection guidelines, and anti-patterns, see modules/phase-structure.md.

Task Format

Each task includes:

  • ID: Unique identifier (TASK-001)
  • Description: Clear action statement
  • Phase: Which phase it belongs to
  • Dependencies: Tasks that must complete first
  • Parallel Marker: [P] if can run concurrently
  • Files: Affected file paths
  • Criteria: How to verify completion

Dependency Rules

Dependencies define execution order and identify parallelization opportunities:

  • Sequential Tasks: Execute in strict order when dependencies exist
  • Parallel Tasks [P]: Can run concurrently when ALL nonconflicting conditions are met
  • File Coordination: Tasks affecting same files MUST run sequentially

Nonconflicting Criteria for Parallel Execution:

  • ✅ Files: No file overlap between tasks
  • ✅ State: No shared configuration or global state
  • ✅ Dependencies: All prerequisites satisfied
  • ✅ Code paths: No merge conflicts possible
  • ✅ Outputs: Tasks don't need each other's results

Mark tasks with [P] ONLY if they pass ALL criteria above.

For fan-out/fan-in patterns, task ID conventions, and validation rules, see modules/dependency-patterns.md.

Example Task Entry

## Phase 2: Core Implementation

### TASK-007 - Implement user authentication service [P]
**Dependencies**: TASK-003, TASK-004
**Files**: src/services/auth.ts, src/types/user.ts
**Criteria**: All auth tests pass, tokens are valid JWT

Verification: Run pytest -v to verify tests pass.

Quality Checklist

  • All requirements mapped to tasks
  • Dependencies are explicit
  • Parallel opportunities identified
  • Tasks are right-sized (not too large/small)
  • Each task has clear completion criteria

Related Skills

  • spec-writing: Creating source specifications
  • speckit-orchestrator: Workflow coordination

Exit Criteria

  • Every task has a unique ID in TASK-NNN format and includes Description, Phase (0-4), Dependencies list, Parallel marker ([P] or sequential), Files, and Criteria fields
  • Tasks marked [P] pass all 5 nonconflicting criteria: no file overlap, no shared state, dependencies satisfied, no merge conflicts possible, no output dependency on each other
  • Every functional requirement from the source spec is mapped to at least one task; unmapped requirements are flagged
  • Quality checklist passes all items before output is finalized
  • Tasks within each phase are ordered so that no task depends on one with a higher task number within the same phase

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Shapes turns action-first: action leads, steps numbered, state restated. Use for ADHD-friendly output. Do not use to trim tokens; use response-compression.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

Inverts burden of proof for code additions. Use when reviewing PRs, planning refactors, or running unbloat to challenge every addition's necessity.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

Tracks per-agent token usage and flags waste in parallel dispatch. Use when evaluating parallel agent efficiency or after a multi-agent run.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

Coordinates Claude agent teams via filesystem protocol. Use when orchestrating parallel agents with task dependencies. Do not use for single-agent tasks.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

Evaluates API surface design, consistency, and exemplar alignment. Use when reviewing public API changes or before releasing a new API surface.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

Selects architecture paradigm via research before scaffolding. Use when architecture is undecided and the choice needs justification and documentation.

The English summary is being prepared. Showing the original description.

athola/claude-night-market343Updated Oct 17, 2026

See all skills by athola

Report a problem with this skill