Agent skill
autoclaude-task-converter
Convert single-file task backlogs (TASKS.md format) to AutoClaude multi-file spec structure. Use when the user wants to (1) convert existing TASKS.md or similar task backlog files to AutoClaude specs, (2) initialize a new project with AutoClaude-compatible task structure, (3) migrate task definitions from simple markdown to the multi-file spec format with requirements.json, context.json, implementation_plan.json. Triggers on phrases like "convert tasks to AutoClaude", "set up AutoClaude specs", "migrate backlog", "create spec from task".
Install this agent skill to your Project
npx add-skill https://github.com/toddward/claude-skills-playground/tree/main/autoclaude-task-converter
SKILL.md
AutoClaude Task Converter
Convert markdown task backlogs to AutoClaude's multi-file spec structure.
Input Format
The skill accepts task files in this markdown format:
### TASK-XXX: Task Title
- **Description**: What needs to be done
- **Area**: Component/Feature area
- **Status**: To Do | In Progress | Done | Blocked
- **Priority**: Critical | High | Medium | Low
- **Dependencies**: TASK-XXX, TASK-YYY (optional)
- **Subtasks**: (optional)
- [ ] Subtask 1
- [x] Subtask 2 (completed)
- **Notes**: Additional context
Also supports single-line tasks: - [ ] Simple task description
Output Structure
For each task, create .auto-claude/specs/XXX-task-name/ containing:
| File | Purpose |
|---|---|
spec.md |
Feature specification |
requirements.json |
Structured requirements |
context.json |
Codebase context (initially empty) |
implementation_plan.json |
Subtasks with status |
Workflow
Converting Existing Tasks
- Locate task file: Find TASKS.md, TODO.md, or similar in project root
- Parse tasks: Extract structured data from markdown
- Create spec directories: Run
scripts/convert_tasks.py - Verify output: Check
.auto-claude/specs/structure
python3 scripts/convert_tasks.py --input TASKS.md --output .auto-claude/specs
New Project Setup
When no task file exists, create the AutoClaude structure:
python3 scripts/convert_tasks.py --init --output .auto-claude/specs
This creates:
.auto-claude/specs/directory.auto-claude/TASKS.mdtemplate file- Sample spec
001-project-setup/
Conversion Rules
Status Mapping
| Input Status | AutoClaude Status |
|---|---|
To Do |
pending |
In Progress |
in_progress |
Done |
completed |
Blocked |
blocked |
Priority Mapping
| Input Priority | AutoClaude Priority |
|---|---|
Critical |
p0 |
High |
p1 |
Medium |
p2 |
Low |
p3 |
Complexity Assessment
Determined by subtask count and description length:
- simple: 0-2 subtasks, single area
- standard: 3-6 subtasks, may span areas
- complex: 7+ subtasks, multiple dependencies
File Schemas
See references/autoclaude_schemas.md for detailed JSON schemas for:
requirements.jsoncontext.jsonimplementation_plan.json
Post-Conversion Note
IMPORTANT: After writing task specs, always inform the user:
NOTE: CURRENTLY AUTOCLAUDE MAY REQUIRE A RESTART BEFORE FULLY UTILIZING TASKS THAT HAVE BEEN WRITTEN OUT.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
toddward/claude-skills-playground
resume-builder
Generate beautiful, professionally designed resumes as 1-2 page PDFs. Creates clean, sophisticated layouts that balance visual appeal with readability, tailored to specific job opportunities when provided.
toddward/claude-skills-playground
family-menu
Creates beautiful, protein-focused weekly dinner menus for families with research capabilities. Generates printer-ready PDFs (8.5x11) with random design styles, identifies leftover opportunities, suggests local restaurants, emphasizes seasonal ingredients, and includes Homemade Pizza Fridays. Use when users want to plan family dinners, create refrigerator menus, or need meal planning assistance.
toddward/claude-skills-playground
thinking-frameworks
Multi-agent thinking frameworks for rigorous problem-solving and decision-making. Contains Dialectical (thesis-antithesis-synthesis), Six Thinking Hats (parallel perspectives), First Principles (strip to fundamentals), and Composition Patterns chaining frameworks together. Use whenever the user needs to evaluate competing approaches, make architecture decisions, challenge assumptions, modernize legacy systems, write RFP responses, run post-mortems, compare technologies, or do any multi-perspective analysis. Triggers on "think this through", "pros and cons", "tradeoffs", "should we use X or Y", "devil's advocate", "argue both sides", "from all angles", "what are we missing", "first principles", "rethink this", "why do we do it this way", "brainstorm", "go/no-go", "use dialectical", "run six hats", or any thinking framework reference.
toddward/claude-skills-playground
brainstorming
Use when creating or developing anything, before writing code or implementation plans - refines rough ideas into fully-formed designs through structured Socratic questioning, alternative exploration, and incremental validation. Optimized for git worktree workflows and Claude CLI agent patterns.
toddward/claude-skills-playground
quarterly-connect
Assist users in creating quarterly connects that act as a strategic partner to guide employees through comprehensive quarterly reflections, helping craft insightful narratives for Quarterly Connection reviews that align with company values and career development goals.
toddward/claude-skills-playground
template-slide-deck
Generate professional slide decks using company-provided template presentations. Use this skill when the user wants to create a presentation based on their own branded PowerPoint template, company slide deck template, or organizational design standards. Triggers include requests like "create slides using our template", "make a presentation with our company format", "generate a deck from this template", or "build slides matching our brand". Also use when the user asks to create slides and has uploaded or mentioned a template .pptx file.
Didn't find tool you were looking for?