aigroup-workflow 1.0.1 → 1.1.1

package/skills/jarvis/claude-simone/legacy/.claude/commands/simone/yolo.md

@@ -1,157 +0,0 @@
-# Execute all open tasks in YOLO mode
-
-This mode is meant to be run without user interaction.
-You **DO NOT** ask the user any questions or ask for confirmation
-If in doubt you **RESEARCH** and **ULTRATHINK** about the best solution.
-
-**IMPORTANT** you try not to stop at any certain point but do so if you need after thoughtful consideration.
-
-Priority is to get the work completed.
-
-## Mode Selection
-
-Check <$ARGUMENTS>:
-
-- If sprint ID provided (e.g., S03): Work ONLY on that sprint
-- If empty: Work on general tasks first, then active sprint tasks
-
-**Report Mode** to the User!
-
-## Safety Guidelines
-
-- **NEVER** modify critical files (.env, alembic migrations, production configs)
-- **STOP** if you encounter database schema changes
-- **STOP** if you need to delete more than 5 files
-- **STOP** if tests are failing after your changes
-- **ALWAYS** run tests after implementing a task
-
-## Follow this exact process
-
-You need to stick to this process and **PRECISELY** follow it
-
-Before you start:
-
-- Run tests to ensure clean baseline using test.md command (@.claude/commands/simone/test.md)
-- **If** FAIL rate is above 10% asses if a fix is possible. If so, fix and move on. If not, move on anyways.
-
-- Check git status to ensure clean working directory
-- **If** git status is not clean just remember and report to the user at the end, but move on.
-
-Also consider <$ARGUMENTS> - if anything between <> can be considered additional instructions, prioritize them in the process.
-
-Get the current datetime stamp from the system and remember it
-
-### FIND OPEN WORK
-
-Execute based on mode:
-
-**If Sprint ID provided in arguments:**
-
-- Navigate to .simone/03_SPRINTS/{sprint_id}/
-- Check if sprint has tasks (T*.md files)
-- If NO tasks exist:
-  - Check for sprint meta file
-  - If meta exists: Jump to ### CREATE SPRINT TASKS
-  - If no meta: Exit with error - sprint doesn't exist
-- If tasks exist: Continue to **Task selection** section below
-
-**If NO arguments (general mode):**
-
-- Use PARALLEL SUBAGENTS to check:
-  - .simone/04_GENERAL_TASKS for open general tasks
-  - .simone/00_PROJECT_MANIFEST.md for currently active sprint
-  - .simone/03_SPRINTS/ for any sprint with only meta file
-- Priority order:
-  1. General tasks (if any open)
-  2. Active sprint tasks (if any open)
-  3. Sprint needing task creation (if found) - maintain order of Sprints
-
-**Task Selection:**
-
-- From found tasks, select ONE that is not completed. Take the lowest ID in Sprint or General Tasks
-- Skip tasks you've previously attempted (check OUTPUT LOG)
-- If no suitable task found and no sprint needs tasks: Exit gracefully
-
-### CREATE SPRINT TASKS
-
-**ONLY EXECUTE** if sprint needs task creation
-
-- Use a **SUBAGENT** and have it include @.claude/commands/simone/create_sprint_tasks.md with Sprint ID as argument
-- Wait for completion
-- After task creation move back to `### FIND OPEN WORK`
-
-### WORK ON TASK
-
-- if you have touched this task before ignore it and jump to the next task
-- if you can't find a task that you have not tried fixing before jump to ### EXECUTE PROJECT REVIEW
-- if you find a task that you cannot fix because the work was done already, close the task and note in Output Log of task.
-- **BEFORE STARTING**: Create a git branch for the task: `git checkout -b task/<task-id>`
-- **USE A SUBAGENT** and have it include @.claude/commands/simone/do_task.md with the Task ID as Argument to execute the Task.
-- **AFTER TASK COMPLETION**: Run tests to verify nothing broke using test.md command (@.claude/commands/simone/test.md)
-- on any failure in the task execution assess the severity of the error:
-  - CRITICAL errors (breaking tests, security issues, data loss risk): **FIX PROBLEMS**
-  - NON-CRITICAL errors (linting, formatting, minor issues): note in OUTPUT LOG and continue
-- on success move on
-
-### COMMIT WORK
-
-- **ONLY IF** tests are passing and no critical issues exist
-- **USE A SUBAGENT** and have it include @.claude/commands/simone/commit.md with the Task ID as Argument and YOLO as additional argument
-- on any failure when committing, note the problem in the OUTPUT LOG of the task and continue
-- after successful commit, merge to main: `git checkout main && git merge task/<task-id>`
-- on success move on
-
-### REPEAT FOR ALL OPEN TASKS
-
-**Based on current mode:**
-
-**Sprint-specific mode:**
-
-- Continue with next task in the specified sprint only
-- Go back to ### WORK ON TASK until all sprint tasks are done
-- When sprint is complete, move to project review
-
-**General mode:**
-
-- After completing a general task, check for more general tasks first
-- If no general tasks remain, move to active sprint tasks
-- Continue until all accessible tasks are done
-- When done move on
-
-## EXECUTE PROJECT REVIEW
-
-- **USE A SUBAGENT** and have it include @.claude/commands/simone/project_review.md
-- Depending on the results of the review:
-  - On FAIL: Think about possible fixes.
-  - If fixes are quickly done, fix right away and repeat `## EXECUTE PROJECT REVIEW``
-  - If fixes are more complex **USE A SUBAGENT** and have it include @.claude/commands/simone/create_general_task.md to create new general tasks as needed.
-  - Go back to `### FIND OPEN WORK` to work on these fixes
-  - On PASS: move on
-
-## MORE WORK TO DO?
-
-**Based on current mode:**
-
-**Sprint-specific mode:**
-
-Your work is done. Move to `## CREATE SUMMARY`
-
-**General mode:**
-
-- Move on to check for more open Tasks or Sprints. Move to `### FIND OPEN WORK`
-
-## CREATE SUMMARY
-
-- Get current datetime stamp from the system and compare to initially remembered timestamp. Calculate duration of the process.
-- Create a summary report including:
-  - Mode executed (Sprint-specific or General)
-  - Sprint tasks created (if applicable)
-  - Number of tasks completed
-  - Number of tasks skipped/failed
-  - Total duration
-  - Any critical issues encountered
-  - Current test status
-  - Next recommended action
-- Report the result of the review to the user including the duration.
-
-Your work is done. Thank you.

package/skills/jarvis/claude-simone/legacy/CHANGELOG.md

@@ -1,71 +0,0 @@
-# Changelog
-
-## 2025-05-30
-
-### Added
-
-- `npx hello-simone` quick start command for easy Simone installation
-- Interactive, adaptive initialize command that:
-  - Auto-detects project type and framework
-  - Guides users through conversational setup
-  - Creates documentation through Q&A process
-  - Adapts to existing vs new projects
-
-### Changed
-
-- initialize command completely rewritten for better user experience
-- Removed complex branching logic in favor of adaptive process
-- Focus on conversational interaction rather than rigid steps
-
-### Improved
-
-- Better handling of existing documentation during setup
-- Smart project detection (Node.js, Python, PHP, etc.)
-- Milestone creation now interactive and context-aware
-
-## 2025-05-29
-
-### Added
-
-- ADR template for architecture decision records
-- Task ID filtering in commit command (T01_S02, TX003 patterns)
-- YOLO mode in commit command to skip user approval
-- Code review results now write to task Output Log sections
-- create_general_task command for structured task creation
-- create_sprint_tasks command for detailed sprint planning
-- create_sprints_from_milestone command for milestone-based sprint planning
-- prime command for quick project context loading
-- yolo command for autonomous task execution without user interaction
-
-### Changed
-
-- Date format simplified to YYYY-MM-DD HH:MM in templates
-- Milestone template heading structure updated
-- do_task command now uses parallel subagents and updates project manifest
-- project_review command removes timeline pressure and focuses on current state
-
-### Removed
-
-- create_sprint, create_task, plan_milestone commands (replaced by new workflow)
-
-## Developer Notes
-
-This update represents a significant evolution based on real-world usage and user feedback. The command architecture has been substantially enhanced with better use of parallel agents, loops, and conditionals, which work surprisingly well in practice.
-
-**Command Complexity**: The enhanced commands may behave differently between Claude Opus and Sonnet models. While the complexity might be challenging for Sonnet, testing suggests it should handle the workflows effectively.
-
-**Task Quality Improvements**: The new `create_sprint_tasks` and `create_general_task` commands now provide much better context and codebase references, resulting in notably higher quality task generation with specific implementation guidance.
-
-**YOLO Mode Warning**: The standalone `yolo` command requires extreme caution. It should only be used:
-
-- Within isolated development environments
-- With Claude Code's permission-skipping mode enabled (`claude --dangerously-skip-permissions`)
-- Never on production systems or systems with important data
-
-The YOLO mode can potentially modify or delete files outside your project directory. Use at your own risk and only in completely isolated environments.
-
-**Feedback Welcome**: Submit issues or pull requests on GitHub. This framework continues to evolve based on real-world usage patterns.
-
-## 2025-05-23
-
-Initial release of Simone framework

package/skills/jarvis/claude-simone/legacy/README.md

@@ -1,219 +0,0 @@
-# Simone for Claude Code (Legacy Version)
-
-## What is this?
-
-Simone is a directory-based project management system I built to work better with Claude Code. It's basically a set of folders, markdown files, and custom commands that help break down software projects into manageable chunks that AI can handle effectively.
-
-**⚠️ Complexity Warning**: Simone is a sophisticated system that requires time to understand properly. It's not a simple plug-and-play solution, but rather a framework that works best when you take the time to learn how it operates and adapt it to your workflow.
-
-**📋 Latest Updates**: See [CHANGELOG.md](CHANGELOG.md) for recent changes and improvements.
-
-## How to Get Started
-
-### 1. Install Simone
-
-```bash
-npx hello-simone
-```
-
-This sets up the folder structure and installs/updates the command files in your project. Can also be used to update an existing installation - command files get backed up automatically.
-
-### 2. Initialize Your Project
-
-Open your project in Claude Code and run:
-
-```
-/simone:initialize
-```
-
-This guides you through the basic setup process. Works with new or existing codebases, and can help you create project documentation (PRDs, architecture docs) or work with documents you already have.
-
-### 3. Set Up Your First Milestone
-
-Create a milestone folder in `.simone/02_REQUIREMENTS/` named `M01_Your_Milestone_Name` (e.g., `M01_Basic_Application`). Include at least:
-
-- `M01_PRD.md` - Product requirements document
-- Other specs as needed: `M01_Database_Schema.md`, `M01_API_Specs.md`, etc.
-
-*Note: There's no command for this yet. Use the existing chat from step 2 to guide Claude through milestone creation, ensuring proper naming with the `M##_` prefix and underscores.*
-
-### 4. Break Down into Sprints
-
-```
-/simone:create_sprints_from_milestone
-```
-
-This analyzes your milestone and breaks it down into logical sprints. It looks at the entire scope and creates meaningful sprint boundaries without detailed tasks yet.
-
-### 5. Create Your First Tasks
-
-```
-/simone:create_sprint_tasks
-```
-
-This analyzes your sprints, reviews documentation, researches necessary information, and identifies knowledge gaps to gain comprehensive understanding of your project. Creates detailed, actionable tasks for the current sprint.
-
-*Important: Only create tasks for your next sprint, not all sprints upfront. After completing Sprint 1, then create tasks for Sprint 2. This ensures the system can reference your existing codebase and incorporate completed work into future task creation.*
-
-### 6. Start Working
-
-```
-/simone:do_task
-```
-
-This will automatically pick a task from your general tasks or sprints. For faster execution, specify a task ID:
-
-```
-/simone:do_task T01_S01
-```
-
-Claude will then work through the specified task with full project context.
-
-That's the basic workflow to get started! You can also:
-
-- Create general tasks with `/simone:create_general_task`
-- Use YOLO mode to run a full sprint autonomously
-- Explore other commands in `.claude/commands/simone/`
-
-**Important**: Simone is a complex system, not a simple set-and-forget tool. It works best when you understand how it operates. Take time to read through the commands and consider adapting them to your workflow.
-
-## How it Works
-
-Simone organizes your project into:
-
-- **Milestones**: Major features or project phases
-- **Sprints**: Groups of related tasks within a milestone
-- **Tasks**: Individual work items scoped for one Claude session
-
-Each task gets full project context so Claude knows exactly what to build and how it fits into your architecture.
-
-## Why I built this
-
-AI coding tools have become incredibly powerful, but they all face the same fundamental challenge: context management. The context window is limited in size, and we have little control over what stays in context and what doesn't.
-
-The problem with long-running sessions is context decay - as you work, critical project knowledge silently falls off the end of the context window. You don't know what's been forgotten until something goes wrong.
-
-My solution: Start fresh for each task, but provide rich surrounding context. By keeping tasks focused and well-scoped, I can dedicate more of the context window to relevant project knowledge, requirements, and architectural decisions. This way:
-
-- Each task starts with exactly the project context it needs
-- No critical knowledge gets lost in long sessions
-- Claude can work confidently with full awareness of requirements
-- The surrounding context guides development, not just the task description
-
-The result is a task-based workflow where Claude always has the right context for the job at hand.
-
-## Key Components
-
-### 00_PROJECT_MANIFEST.md
-
-The central document containing the project's vision, goals, and high-level overview. It serves as the starting point for Claude to understand the project. **Important**: The manifest file must be named exactly `00_PROJECT_MANIFEST.md`, not `MANIFEST.md`.
-
-### 01_PROJECT_DOCS/
-
-Contains general project documentation including technical specifications, user guides, and API documentation that Claude can reference.
-
-### 02_REQUIREMENTS/
-
-Organized by milestones, this directory stores product requirements documents (PRDs) and their amendments, providing a clear view of what needs to be built. This helps Claude understand the project requirements. Milestone folders must follow the naming convention `M##_Milestone_Name/` (e.g., `M01_Backend_Setup/`).
-
-### 03_SPRINTS/
-
-Contains sprint plans and task definitions organized by milestone and sprint sequence. Each sprint folder contains individual task files with detailed information for Claude to work on.
-
-### 04_GENERAL_TASKS/
-
-Stores task definitions for work not tied to a specific sprint. Completed tasks use a `TX` prefix (e.g., `TX001_Completed_Task.md`), making it easy for Claude to identify what's been done.
-
-### 05_ARCHITECTURAL_DECISIONS/
-
-Captures significant architectural decisions as Architecture Decision Records (ADRs), documenting the context, options considered, and rationale. This provides critical context for Claude when making technical decisions. Uses a structured ADR template for consistency.
-
-### 10_STATE_OF_PROJECT/
-
-Contains timestamped project review snapshots created by the `project_review` command. These provide a historical record of project health, technical decisions, and progress over time.
-
-### 99_TEMPLATES/
-
-Contains standardized templates for different document types to ensure consistency for both humans and Claude:
-
-- Task templates with structured objectives and acceptance criteria
-- Sprint and milestone metadata templates
-- ADR template for documenting architectural decisions
-- All templates use simplified date formats (YYYY-MM-DD HH:MM)
-
-### .claude/commands/simone/
-
-Custom Claude Code commands that power the Simone workflow:
-
-- `initialize` - Set up project structure and documentation
-- `create_sprints_from_milestone` - Break milestones into logical sprints
-- `create_sprint_tasks` - Generate detailed tasks from sprint plans
-- `do_task` - Execute individual tasks with full context
-- `yolo` - Autonomous sprint execution (use with caution)
-- And many more for testing, reviewing, and project management
-
-## Directory Structure
-
-```plaintext
-.simone/
-├── 00_PROJECT_MANIFEST.md
-├── 01_PROJECT_DOCS/
-├── 02_REQUIREMENTS/
-│   ├── M01_Backend_Setup/
-│   ├── M02_Frontend_Setup/
-│   └── ...
-├── 03_SPRINTS/
-│   ├── S01_M01_Initial_API/
-│   ├── S02_M01_Database_Schema/
-│   └── ...
-├── 04_GENERAL_TASKS/
-│   ├── TX001_Refactor_Logging_Module.md  # Completed task
-│   ├── T002_API_Rate_Limiting.md          # Open task
-│   └── ...
-├── 05_ARCHITECTURE_DECISIONS/
-│   ├── ADR001_Database_Selection.md
-│   └── ...
-├── 10_STATE_OF_PROJECT/         # Project review snapshots
-└── 99_TEMPLATES/
-    ├── task_template.md
-    ├── sprint_meta_template.md
-    └── milestone_meta_template.md
-```
-
-## Configuration Tips
-
-### Enabling Parallel Task Execution
-
-While Simone commands like `create_sprint_tasks` support the `useParallelSubagents` instruction, Claude Code needs to be configured to actually execute tasks in parallel. By default, it only runs one task at a time.
-
-To enable parallel execution:
-
-```bash
-# Set the number of parallel tasks (example: 3)
-claude config set --global "parallelTasksCount" 3
-
-# Check your current configuration
-claude config list -g
-```
-
-**Important considerations:**
-
-- Choose the number based on your system's capabilities and rate limits
-- Parallel execution increases API usage significantly
-- Some tasks may have conflicts when run in parallel
-- Start with a small number (2-3) and adjust based on your experience
-
-## Contributing & Feedback
-
-I'd love to hear from you! This is very much shaped by how I work, and I'm sure there are tons of improvements to be made.
-
-- **GitHub Issues**: Best place for bugs and feature requests
-- **Anthropic Discord**: Find me @helmi if you want to chat about it
-- **Pull Requests**: Very welcome! Let's make this better together
-
-I'm particularly interested in:
-
-- How you're using it differently
-- What's missing for your workflow
-- Ideas for better Claude Code integration
-- Different organizational approaches