Release v1.0.0: Design Thinking Enabler with advanced features

Double Diamond Methodology:
- probe (Discover), grasp (Define), tangle (Develop), ink (Deliver)
- embrace: Full 4-phase workflow with session checkpoints

Autonomy Modes (-a flag):
- autonomous, semi-autonomous (default), supervised, loop-until-approved

Session Recovery (-R flag):
- Checkpoint after each phase, resume interrupted workflows

Loop-Until-Approved (-l flag):
- Retry failed tasks until quality gate passes (configurable threshold)

Specialized Agent Roles:
- Architect, Researcher, Reviewer, Implementer, Synthesizer

Nano Banana Prompt Refinement:
- Intelligent image prompt enhancement
- Type-specific optimization (app-icon, social-media, diagram)

Enhanced Image Auto-Routing:
- Detects app icons, favicons, diagrams, social media banners
- Hero images, thumbnails, marketing graphics

Quality Gates:
- Configurable threshold (-q flag, default 75%)
- Status: PASSED (>=90%), WARNING (75-89%), FAILED (<75%)

Octopus Humor:
- ASCII art mascot throughout codebase
- Witty phase banners and documentation

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: nyldn

.gitignore

@@ -1,5 +1,8 @@
 # Claude Octopus - Git Ignore
 
+# Claude Code
+CLAUDE.md
+
 # macOS
 .DS_Store
 .AppleDouble

CHANGELOG.md

@@ -0,0 +1,101 @@
+# Changelog
+
+All notable changes to Claude Octopus will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [1.0.0] - 2026-01-15
+
+### Added
+
+#### Double Diamond Methodology
+Design Thinking workflow with octopus-themed commands:
+- **probe** - Parallel research from 4 perspectives with AI synthesis (Discover phase)
+- **grasp** - Multi-agent consensus building on problem definition (Define phase)
+- **tangle** - Enhanced map-reduce with quality gates (Develop phase)
+- **ink** - Validation and final deliverable generation (Deliver phase)
+- **embrace** - Full 4-phase Double Diamond workflow
+- **preflight** - Dependency validation before workflows
+
+#### Intelligent Auto-Routing
+Smart task classification routes to appropriate agents:
+- **Image generation**: App icons, favicons, diagrams, social media banners, hero images
+- **Code review**: Security audits, code analysis, PR reviews
+- **Coding**: Implementation, debugging, refactoring
+- **Design**: UI/UX analysis, accessibility, component design
+- **Research**: Documentation, architecture analysis, best practices
+- **Copywriting**: Marketing copy, content generation
+
+#### Nano Banana Prompt Refinement
+Intelligent prompt enhancement for image generation:
+- Automatic detection of image type (app-icon, social-media, diagram, general)
+- Type-specific prompt optimization for better visual results
+- Integrated into auto-routing for seamless UX
+
+#### Autonomy Modes
+Configurable human oversight levels:
+- **autonomous** - Full auto, proceed on failures
+- **semi-autonomous** - Pause on quality gate failures (default)
+- **supervised** - Human approval after each phase
+- **loop-until-approved** - Retry failed tasks until quality gate passes
+
+#### Session Recovery
+Resume interrupted workflows:
+- Automatic checkpoint after each phase completion
+- Session state persisted to JSON
+- Resume with `-R` flag from last successful phase
+
+#### Specialized Agent Roles
+Role-based agent selection for phases:
+- **Architect** - System design and planning (Codex Max)
+- **Researcher** - Deep investigation (Gemini Pro)
+- **Reviewer** - Code review and validation (Codex Review)
+- **Implementer** - Code generation (Codex Max)
+- **Synthesizer** - Result aggregation (Gemini Flash)
+
+#### Quality Gates
+Configurable quality thresholds:
+- Default 75% success threshold (configurable with `-q`)
+- Quality gate status: PASSED (>=90%), WARNING (75-89%), FAILED (<75%)
+- Loop-until-approved retry logic (up to 3 retries by default)
+
+#### Multi-Agent Orchestration
+Core execution patterns:
+- **spawn** - Single agent execution
+- **fan-out** - Same prompt to all agents
+- **map-reduce** - Task decomposition and parallel execution
+- **parallel** - JSON-defined task execution
+
+#### Agent Fleet
+Premium model defaults (Jan 2026):
+- `codex` - GPT-5.1-Codex-Max (premium default)
+- `codex-standard` - GPT-5.2-Codex
+- `codex-max` - GPT-5.1-Codex-Max
+- `codex-mini` - GPT-5.1-Codex-Mini
+- `codex-general` - GPT-5.2
+- `gemini` - Gemini-3-Pro-Preview
+- `gemini-fast` - Gemini-3-Flash-Preview
+- `gemini-image` - Gemini-3-Pro-Image-Preview
+- `codex-review` - GPT-5.2-Codex (review mode)
+
+#### CLI Options
+- `-p, --parallel NUM` - Max parallel agents (default: 3)
+- `-t, --timeout SECS` - Timeout per task (default: 300s)
+- `-a, --autonomy MODE` - Set autonomy mode
+- `-q, --quality NUM` - Quality gate threshold
+- `-l, --loop` - Enable loop-until-approved
+- `-R, --resume` - Resume interrupted session
+- `-v, --verbose` - Verbose output
+- `-n, --dry-run` - Show what would be done
+
+#### Documentation
+- Comprehensive README with Double Diamond methodology
+- Octopus Philosophy section explaining the metaphor
+- Troubleshooting guide with witty octopus tips
+- ASCII art mascot throughout codebase
+
+### Notes
+
+Initial release as a Claude Code plugin for Design Thinking workflows.
+Built with multi-agent orchestration using Codex CLI and Gemini CLI.