Expertise in breaking down specifications into executable, dependency-ordered tasks. Activates when user discusses task planning, work breakdown, or implementation ordering. Trigger keywords: task decomposition, work breakdown, task list, dependency order, implementation tasks, tasks.md, T001, parallelizable
Breaks down specifications into executable, dependency-ordered tasks for developers to implement.
/plugin marketplace add datamaker-kr/synapse-claude-marketplace/plugin install speckit-helper@synapse-marketplaceThis skill is limited to using the following tools:
references/dependency-patterns.mdreferences/phase-ordering.mdreferences/task-format.mdThis skill provides expertise in decomposing specifications into executable, dependency-ordered implementation tasks. It transforms structured requirements and user stories into a concrete work plan that developers can follow sequentially or in parallel. The output is a tasks.md file with numbered tasks, dependency annotations, and full traceability back to the specification.
The skill is triggered when the conversation involves:
tasks.md fileTasks are organized into five sequential phases:
| Phase | Purpose | Examples |
|---|---|---|
| Setup | Project scaffolding, tooling, configuration | Init repo, install deps, configure linter |
| Foundation | Core infrastructure and data layer | Database schema, base models, auth setup |
| Stories | Feature implementation per user story | Implement US1, US2, US3 endpoints and UI |
| Integration | Cross-feature wiring, E2E flows | API integration, state management, routing |
| Finalization | Quality assurance, polish, deployment prep | Testing, docs, CI/CD, performance tuning |
Tasks within each phase are ordered by dependency. A later phase never starts before its prerequisites in earlier phases are complete.
Every task follows this standardized format:
- [ ] T001 [P] [US1] path/to/file.ts -- Description (S) [Spec FR-001]
Where:
T001 -- Unique task identifier, zero-padded and sequential[P] -- Parallelizable marker (present only if the task can run concurrently with others)[US1] -- User story reference linking to the specificationpath/to/file.ts -- Primary file or directory affectedDescription -- Concise action description starting with a verb(S) -- Size estimate: (S) small, (M) medium, (L) large[Spec FR-001] -- Traceability reference back to a specific requirement sectionThe skill analyzes tasks to detect and annotate dependencies:
Dependencies are expressed as depends: T001, T003 annotations when non-obvious ordering exists.
Tasks that share no dependencies are marked with the [P] flag, indicating they can be executed concurrently. The skill groups parallelizable tasks together within each phase to maximize throughput.
After generating the task list, the skill validates coverage:
FR-XXX in the specification maps to at least one taskNFR-XXX has a corresponding task or is addressed by a cross-cutting taskUS1, US2, ...) appears in at least one task's [USx] tagWhen coverage validation detects missing mappings, the skill inserts explicit gap markers:
- [ ] T015 [Gap] -- No task covers FR-012 (payment retry logic) [Spec FR-012]
Gap markers are clearly labeled with [Gap] so they can be identified and resolved before implementation begins.
The decomposition process follows these steps:
spec.md to extract all FR, NFR, user stories, data model entities, and API contracts.[P].[Gap] markers for any uncovered requirements.For detailed format specifications, dependency patterns, and phase ordering rules, consult:
references/task-format.md -- Full task format specification with examplesreferences/dependency-patterns.md -- Common dependency patterns and resolution strategiesreferences/phase-ordering.md -- Phase definitions, transition criteria, and ordering rulesActivates when the user asks about AI prompts, needs prompt templates, wants to search for prompts, or mentions prompts.chat. Use for discovering, retrieving, and improving prompts.
Search, retrieve, and install Agent Skills from the prompts.chat registry using MCP tools. Use when the user asks to find skills, browse skill catalogs, install a skill for Claude, or extend Claude's capabilities with reusable AI agent components.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.