Reusable AI agent configurations for development workflows. Designed for XP/TDD practitioners who want consistent, high-quality AI assistance.
.
├── .agents/rules/
│ └── base.md # 📌 Single source of truth for all AI rules
├── .cursor/
│ ├── commands/ # Slash commands (Cursor IDE)
│ │ ├── fic-*.md # FIC workflow commands
│ │ ├── xp-*.md # XP/TDD commands
│ │ └── eb-*.md # Eventbrite-specific commands
│ └── rules/ # Cursor rules
│ ├── base.mdc # Core XP/TDD principles
│ ├── fic-workflow.mdc # FIC context management
│ └── *.mdc # Specialized rules
├── src/thoughts/ # Node/TS CLI for thoughts/ management
├── thoughts/ # Research docs and plans (FIC workflow)
├── AGENTS.md → base.md # Symlink for OpenAI Codex
├── CLAUDE.md → base.md # Symlink for Claude
└── GEMINI.md → base.md # Symlink for Gemini
Based on stepwise-dev and the FIC methodology.
Problem: LLMs lose attention after ~60% context usage.
Solution: Structured phases with intentional context clearing:
📖 Research → Save to thoughts/ → Clear context
📋 Plan → Save to thoughts/ → Clear context
⚙️ Implement (phase by phase) → Clear between phases
✅ Validate → Report
| Command | Purpose |
|---|---|
/fic-research |
Document codebase comprehensively, save to thoughts/shared/research/ |
/fic-create-plan |
Create detailed implementation plans iteratively |
/fic-implement-plan |
Execute plans phase by phase with verification |
/fic-validate-plan |
Verify implementation against plan |
Persistent storage for research and plans (tracked in git):
thoughts/
├── {username}/ # Personal notes (you write)
│ ├── tickets/
│ └── notes/
├── shared/ # Team-shared (AI writes, tracked in git)
│ ├── research/ # Research documents (e.g., auto-improvement mechanisms)
│ ├── plans/ # Implementation plans (e.g., feedback loop activation)
│ └── prs/ # PR descriptions
└── searchable/ # Hardlinks for fast grep (gitignored)
Research documents and implementation plans in thoughts/shared/ are committed to the repository to maintain project knowledge and enable collaboration.
Node/TS CLI for managing thoughts/:
cd src/thoughts
npm install
npm run build
# Commands
npx thoughts init # Initialize thoughts/ structure
npx thoughts sync # Sync hardlinks after adding files
npx thoughts metadata # Get git metadata for frontmatterXP/TDD commands use the xp- prefix. Eventbrite-specific variants use the eb- prefix.
| Command | Purpose |
|---|---|
/xp-code-review |
Review pending changes (tests, maintainability, rules) |
/xp-increase-coverage |
Identify and test high-value untested code |
/xp-plan-untested-code |
Create actionable plan to cover untested code |
/xp-predict-problems |
Predict likely production failures |
/xp-mikado-method |
Guide safe, incremental refactoring |
/xp-technical-debt |
Catalog and prioritize technical debt |
/xp-refactor |
Apply XP Simple Design principles |
/xp-simple-design-refactor |
Maintainability & Simple Design refactoring |
/xp-security-analysis |
Pragmatic security risk analysis |
Eventbrite-specific commands use the eb- prefix to distinguish them from generic XP/TDD commands.
| Command | Purpose |
|---|---|
/eb-code-review |
Eventbrite-specific code review variant |
/eb-increase-coverage |
Eventbrite-specific coverage analysis |
/eb-plan-untested-code |
Eventbrite-specific untested code planning |
/eb-predict-problems |
Eventbrite-specific problem prediction |
/eb-mikado-method |
Eventbrite-specific Mikado method variant |
/eb-technical-debt |
Eventbrite-specific technical debt catalog |
/eb-security-analysis |
Eventbrite-specific security analysis |
/eb-bug-fixing-agent |
Eventbrite bug fixing expert with OWASP, threat modeling, cloud security |
| Rule | Purpose | Activation |
|---|---|---|
base.mdc |
Core XP/TDD principles | Always active |
ai-feedback-learning-loop.mdc |
AI feedback and rule refinement cycle | Always active |
fic-workflow.mdc |
FIC context management | Manual |
tdd-workflow.mdc |
TDD-specific rules | Manual |
refactoring.mdc |
Safe refactoring | Manual |
debugging.mdc |
Systematic debugging | Manual |
python-dev.mdc |
Python-specific | Auto on *.py |
# Automated sync (recommended)
cd ~/saski/augmentedcode-configuration
./sync-cursor-config.sh repo-to-global
# Or manual copy
cp -r .cursor/commands/* ~/.cursor/commands/
cp -r .cursor/rules/* ~/.cursor/rules/
# Restart CursorThis repository maintains bidirectional sync with your global Cursor configuration (~/.cursor/). Use the sync script to keep them synchronized:
cd ~/saski/augmentedcode-configuration
# Sync both directions (default)
./sync-cursor-config.sh
# Or specify direction
./sync-cursor-config.sh repo-to-global # Repository → Global
./sync-cursor-config.sh global-to-repo # Global → RepositorySee .cursor/rules/cursor-config-management.mdc for detailed sync workflow.
# Copy to your project
cp -r .cursor /path/to/your/project/# Symlink for Claude Code
ln -s .agents/rules/base.md CLAUDE.mdThese configurations enforce:
- TDD: Test-first, one failing test at a time
- Baby Steps: Small, incremental changes
- Simple Design: Clarity over cleverness
- High Quality: Strict validation before commits
- Context Engineering: Manage AI context effectively
Unlicense — Public Domain