@arvorco/relentless 0.1.0

package/CHANGES_SUMMARY.md

@@ -0,0 +1,255 @@
+# Relentless Commands Refactoring - Complete Summary
+
+## ✅ What Was Accomplished
+
+Successfully forked Spec Kit commands into native Relentless commands with multi-agent support strategy.
+
+## 🎯 Key Changes
+
+### 1. Command Renaming (Spec Kit → Relentless)
+- ✅ All `/speckit.*` commands renamed to `/relentless.*`
+- ✅ 9 commands updated: constitution, specify, plan, tasks, checklist, clarify, analyze, implement, taskstoissues
+- ✅ All internal references updated
+
+### 2. Skills Architecture Implementation
+**Created 9 new comprehensive skills:**
+```
+.claude/skills/
+├── constitution/    # Personalized project governance
+├── specify/         # Feature specification
+├── plan/            # Technical implementation plan
+├── tasks/           # User stories & tasks (PRD source!)
+├── checklist/       # Quality validation
+├── clarify/         # Ambiguity resolution
+├── analyze/         # Consistency checking
+├── implement/       # Implementation workflow
+└── taskstoissues/   # GitHub issues generation
+```
+
+**Each skill includes:**
+- Comprehensive SKILL.md with step-by-step instructions
+- Templates (moved from `.specify/templates/`)
+- Scripts (refactored from `.specify/scripts/`)
+
+**Commands are now thin wrappers** (~10-20 lines) that load skills.
+
+### 3. Directory Structure Migration
+**FROM:**
+```
+.specify/specs/NNN-feature/
+├── spec.md
+├── plan.md
+└── ...
+```
+
+**TO:**
+```
+relentless/features/NNN-feature/
+├── spec.md          # Feature description
+├── plan.md          # Technical plan
+├── tasks.md         # USER STORIES (source for prd.json!)
+├── checklist.md     # Quality checks
+├── prd.json         # For orchestrator
+└── progress.txt     # Progress tracking
+```
+
+### 4. Constitution Flow Fixed
+**OLD (Wrong):**
+- `relentless init` → copies template to `relentless/constitution.md` ❌
+
+**NEW (Correct):**
+- `relentless init` → creates empty structure
+- `/relentless.constitution` → asks questions, generates personalized constitution ✓
+
+Each project gets a **personalized** constitution, not a copied template.
+
+### 5. Multi-Agent Support Strategy
+
+#### Tier 1: Full Skills Support ✅
+**Agents:** Claude Code, Amp, OpenCode
+- ✅ Complete `/relentless.*` command support
+- ✅ Interactive workflows
+- ✅ Best user experience
+
+#### Tier 2: Extensions/Hybrid 🔄
+**Agents:** Gemini
+- 🔄 Extensions support (in development)
+- 📄 [GEMINI_SETUP.md](./GEMINI_SETUP.md) created with guide
+- ✅ Workaround: Manual file creation + prompting
+
+#### Tier 3: Manual/CLI 📝
+**Agents:** Droid, Codex
+- ✅ Manual file creation workflow
+- ✅ Can prompt agent to create files
+- ✅ Can reference skill files for instructions
+- ✅ All CLI commands work
+
+### 6. Updated Workflow
+
+**New Complete Workflow:**
+```bash
+# 1. Initialize
+relentless init
+
+# 2. Create personalized constitution
+/relentless.constitution
+
+# 3. Create feature spec
+/relentless.specify Add user authentication
+
+# 4. Generate plan, tasks, checklist
+/relentless.plan
+/relentless.tasks
+/relentless.checklist
+
+# 5. Convert tasks.md to prd.json
+relentless convert relentless/features/003-user-auth/tasks.md --feature 003-user-auth
+
+# 6. Run orchestration
+relentless run --feature 003-user-auth --tui
+```
+
+### 7. Convert Command Enhanced
+**OLD:**
+- Read any markdown file
+- Assumed `prd.md` format
+
+**NEW:**
+- Reads `tasks.md` (primary source for user stories)
+- Optional `--with-checklist` flag
+- TODO: Merge checklist criteria into acceptance criteria
+- Future: Read spec.md/plan.md for additional context
+
+### 8. Scaffolder Updates
+**Init now installs 11 skills:**
+- ✅ constitution (new)
+- ✅ specify (new)
+- ✅ plan (new)
+- ✅ tasks (new)
+- ✅ checklist (new)
+- ✅ clarify (new)
+- ✅ analyze (new)
+- ✅ implement (new)
+- ✅ taskstoissues (new)
+- ✅ prd (existing)
+- ✅ relentless (existing)
+
+**Does NOT** copy constitution.md - must be generated via `/relentless.constitution`
+
+### 9. Documentation Updated
+- ✅ README.md - Updated workflows, agent support tiers, quick start
+- ✅ REFACTOR_SUMMARY.md - Complete technical summary
+- ✅ GEMINI_SETUP.md - Gemini extensions guide
+- ✅ CHANGES_SUMMARY.md - This file!
+
+## 📁 Files Modified
+
+### Deleted (speckit commands)
+```
+.claude/commands/speckit.*.md (9 files)
+```
+
+### Created (new commands)
+```
+.claude/commands/relentless.*.md (9 files)
+```
+
+### Created (new skills)
+```
+.claude/skills/constitution/
+.claude/skills/specify/
+.claude/skills/plan/
+.claude/skills/tasks/
+.claude/skills/checklist/
+.claude/skills/clarify/
+.claude/skills/analyze/
+.claude/skills/implement/
+.claude/skills/taskstoissues/
+```
+
+### Modified (core files)
+```
+bin/relentless.ts            # Updated convert command
+src/init/scaffolder.ts       # Updated skill installation
+README.md                    # Updated workflows & agent support
+relentless/features/ghsk-ideas/prd.md  # Updated description
+.specify/scripts/*.sh        # Refactored paths to relentless/features/
+.specify/templates/*.md      # Updated references
+```
+
+## 🎨 What Makes This Special
+
+1. **Agent-Agnostic Core**: All agents can use `relentless run` for orchestration
+2. **Graceful Degradation**: Best experience for Claude/Amp/OpenCode, still works for everyone
+3. **Extensible**: Easy to add support for new agents
+4. **Skills Own Templates**: Self-contained, easy to maintain
+5. **Thin Command Layer**: Commands are simple wrappers, logic in skills
+6. **Personalized Governance**: Constitution generated per-project
+7. **Clear Source of Truth**: tasks.md → prd.json (user stories)
+
+## 🚀 What Users Get
+
+### Claude Code/Amp/OpenCode Users
+- Interactive `/relentless.*` commands
+- Guided workflows with validation
+- Template-based generation
+- Best experience
+
+### Gemini Users (Coming Soon)
+- Extension-based workflow
+- Similar experience via different mechanism
+- Conversion scripts planned
+
+### Droid/Codex Users
+- Manual file creation
+- Can prompt agent: "Create tasks.md following .claude/skills/tasks/SKILL.md format"
+- All CLI commands work
+- Full orchestration support
+
+## 📝 Next Steps
+
+### High Priority
+- [ ] Test complete workflow with each agent tier
+- [ ] Implement checklist merging logic in convert command
+- [ ] Test end-to-end with real projects
+
+### Medium Priority
+- [ ] Create Gemini extensions conversion script
+- [ ] Add `--with-plan` option to features create
+- [ ] Migration script for old `.specify/` projects
+
+### Low Priority
+- [ ] Video tutorials
+- [ ] Example projects per agent tier
+- [ ] Performance optimization
+
+## 🔗 References
+
+- [REFACTOR_SUMMARY.md](./REFACTOR_SUMMARY.md) - Technical details
+- [GEMINI_SETUP.md](./GEMINI_SETUP.md) - Gemini extensions guide
+- [README.md](./README.md) - User documentation
+- Skills: `.claude/skills/*/SKILL.md` - Implementation guides
+
+## 💡 Key Insights
+
+1. **Option D (Hybrid) was the right choice** - Best for power users, works for everyone
+2. **OpenCode now supports skills** - Can be Tier 1
+3. **Gemini extensions are easy to hybridize** - Worth the effort
+4. **Droid/Codex users can reference skills** - Good workaround
+5. **tasks.md is the source of truth** - Clearer than spec.md
+6. **Constitution must be personalized** - Template copying was wrong
+
+## 🎉 Success Metrics
+
+- ✅ All commands renamed and functional
+- ✅ 9 comprehensive skills created
+- ✅ Scripts refactored to new paths
+- ✅ Multi-agent strategy documented
+- ✅ README fully updated
+- ✅ Backward compatible (legacy PRD workflow still works)
+
+---
+
+**Status:** ✅ Complete and ready for testing
+
+**Next Action:** Test the complete workflow with different agent tiers

package/CLAUDE.md

@@ -0,0 +1,92 @@
+# Relentless - Universal AI Agent Orchestrator
+
+## Overview
+
+This is the Relentless codebase - a universal AI agent orchestrator that works with multiple AI coding agents (Claude Code, Amp, OpenCode, Codex, Droid, Gemini).
+
+**Recent Major Refactoring (January 2026):**
+- Forked GitHub Spec Kit-inspired commands as native Relentless commands
+- All `/speckit.*` commands renamed to `/relentless.*`
+- Implemented skills architecture: commands are thin wrappers, logic in skills
+- Multi-tier agent support: Full (Claude/Amp/OpenCode), Extensions (Gemini), Manual (Droid/Codex)
+- See [REFACTOR_SUMMARY.md](./REFACTOR_SUMMARY.md) and [CHANGES_SUMMARY.md](./CHANGES_SUMMARY.md) for details
+
+## Codebase Patterns
+
+### Directory Structure
+
+- `bin/` - CLI entry point (relentless.ts)
+- `src/` - Core TypeScript implementation
+  - `agents/` - Agent adapters for each AI coding agent
+  - `config/` - Configuration schema and loading
+  - `prd/` - PRD parsing and validation
+  - `execution/` - Orchestration loop and routing
+  - `init/` - Project initialization scaffolder
+- `.claude/skills/` - Skills for Claude Code/Amp/OpenCode
+- `.claude/commands/` - Command wrappers that load skills
+- `templates/` - Templates copied to projects on init
+
+### Key Concepts
+
+1. **Agent Adapters**: Each AI agent has an adapter implementing `AgentAdapter` interface
+2. **Skills Architecture**: Commands load skills which contain the actual logic and templates
+3. **PRD Format**: User stories in `prd.json` with `passes: true/false` status (generated from `tasks.md`)
+4. **Completion Signal**: `<promise>COMPLETE</promise>` indicates all stories done
+5. **Progress Log**: `progress.txt` accumulates learnings across iterations
+6. **Multi-Tier Agent Support**: Full skills support (Claude/Amp/OpenCode), Extensions (Gemini), Manual (Droid/Codex)
+
+### Development Commands
+
+```bash
+# Install locally for development
+bun install
+
+# Run the orchestrator during development
+bun run bin/relentless.ts run --feature <name>
+
+# Type check
+bun run typecheck
+
+# Lint
+bun run lint
+```
+
+### Testing Locally
+
+```bash
+# Test init command
+mkdir /tmp/test && cd /tmp/test
+bun run /path/to/relentless/bin/relentless.ts init
+
+# Test with a simple PRD
+cd /path/to/relentless
+bun run bin/relentless.ts run --feature <name>
+```
+
+### Using the Global Binary
+
+```bash
+# Install globally
+bun install -g .
+
+# Run from anywhere
+relentless init
+relentless run --feature <name>
+```
+
+### Code Style
+
+- TypeScript with strict mode
+- Bun as runtime (no Node.js)
+- Zod for schema validation
+- Commander for CLI parsing
+
+### Important Files
+
+- `bin/relentless.ts` - Main CLI entry point
+- `src/agents/` - Agent adapters for each supported agent
+- `src/init/scaffolder.ts` - Project initialization and skill installation
+- `src/prd/parser.ts` - PRD markdown parser (reads tasks.md format)
+- `.claude/skills/*/SKILL.md` - Skill implementations
+- `.claude/commands/relentless.*.md` - Command wrappers
+- `templates/` - Default templates for constitution, plan, etc.

package/GEMINI_SETUP.md

@@ -0,0 +1,256 @@
+# Gemini Setup for Relentless
+
+## Current Status
+
+🔄 **Gemini support is functional for orchestration but extensions are in development.**
+
+**What works now:**
+- ✅ Gemini agent adapter (`relentless run --agent gemini`)
+- ✅ Orchestration with `relentless run`
+- ✅ Rate limit detection and fallback
+- ✅ All CLI commands work
+
+**What's coming:**
+- 🔄 Extensions for `/relentless.*` commands (Gemini-native format)
+- 🔄 Automatic extension installation
+- 🔄 Skills → Extensions conversion script
+
+## Quick Start (Current)
+
+Use Gemini with the manual workflow:
+
+```bash
+# 1. Install Relentless
+bun install -g github:ArvorCo/Relentless
+cd your-project
+relentless init
+
+# 2. Create files manually or prompt Gemini
+gemini "Create relentless/constitution.md with governance rules for a TypeScript project"
+gemini "Create relentless/features/001-user-auth/spec.md with specification for user authentication"
+gemini "Create relentless/features/001-user-auth/plan.md with technical plan using React/PostgreSQL"
+gemini "Create relentless/features/001-user-auth/tasks.md with user stories in this format:
+### US-001: Title
+**Description:** As a user, I want X so that Y.
+**Acceptance Criteria:**
+- [ ] Criterion 1
+- [ ] Typecheck passes"
+
+# 3. Convert and run orchestration with Gemini
+relentless convert relentless/features/001-user-auth/tasks.md --feature 001-user-auth
+relentless run --feature 001-user-auth --agent gemini --tui
+```
+
+## Using Skill Instructions
+
+Gemini can read and follow the skill instructions:
+
+```bash
+gemini "Read .claude/skills/tasks/SKILL.md and create tasks.md in relentless/features/001-user-auth/ following that format exactly"
+```
+
+## Structure (When Extensions Are Ready)
+
+Gemini extensions will be configured separately from Claude skills:
+
+```
+your-project/
+├── relentless/
+│   ├── constitution.md
+│   └── features/
+├── .claude/
+│   └── skills/              # For Claude/Amp/OpenCode
+└── .gemini/                 # Future: Gemini extensions
+    └── extensions/
+        ├── constitution/
+        ├── specify/
+        ├── plan/
+        ├── tasks/
+        └── checklist/
+```
+
+## Extension Format
+
+Gemini extensions use a different format than Claude skills. Each extension needs:
+
+### Example: constitution Extension
+
+```json
+{
+  "name": "constitution",
+  "description": "Create or update project constitution",
+  "schema": {
+    "type": "object",
+    "properties": {
+      "action": {
+        "type": "string",
+        "enum": ["create", "update"],
+        "description": "Whether to create new or update existing constitution"
+      },
+      "context": {
+        "type": "string",
+        "description": "Additional context about project requirements"
+      }
+    }
+  },
+  "instructions": "Load from SKILL.md content here..."
+}
+```
+
+## Setup Instructions
+
+### Option 1: Automatic Setup (Coming Soon)
+
+```bash
+relentless init --agent gemini
+# This will create .gemini/extensions/ with all Relentless extensions
+```
+
+### Option 2: Manual Setup (Current)
+
+1. **Copy Skills to Gemini Format:**
+   ```bash
+   # Convert skills to Gemini extensions format
+   # (Script coming soon)
+   ```
+
+2. **Register Extensions:**
+   Follow Gemini's extension registration process for your project.
+
+3. **Use Extensions:**
+   ```bash
+   # In Gemini chat:
+   @constitution create constitution for TypeScript project with strict testing
+   @specify Add user authentication with OAuth2
+   @plan using React and PostgreSQL
+   @tasks break down into user stories
+   @checklist generate quality validation
+   ```
+
+## Workflow Comparison
+
+### Claude Code/Amp (Skills)
+```bash
+/relentless.constitution
+/relentless.specify Add user auth
+/relentless.plan
+/relentless.tasks
+/relentless.checklist
+```
+
+### Gemini (Extensions)
+```bash
+@constitution create for TypeScript project
+@specify Add user auth
+@plan using React/PostgreSQL
+@tasks generate stories
+@checklist create validation
+```
+
+## Hybrid Approach
+
+While Gemini extension support is being developed, you can use the manual workflow:
+
+1. **Create feature directory:**
+   ```bash
+   relentless features create user-auth
+   ```
+
+2. **Prompt Gemini to create files:**
+   ```
+   Create relentless/features/001-user-auth/spec.md with specification for:
+   - User authentication
+   - Email/password login
+   - JWT tokens
+   ```
+
+3. **Convert and run:**
+   ```bash
+   relentless convert relentless/features/001-user-auth/tasks.md --feature 001-user-auth
+   relentless run --feature 001-user-auth --agent gemini
+   ```
+
+## Orchestrator Configuration
+
+Gemini is already configured in `relentless/config.json`:
+
+```json
+{
+  "defaultAgent": "claude",
+  "agents": {
+    "gemini": {
+      "dangerouslyAllowAll": true
+    }
+  },
+  "fallback": {
+    "enabled": true,
+    "priority": ["claude", "codex", "amp", "opencode", "gemini"],
+    "autoRecovery": true
+  }
+}
+```
+
+You can set Gemini as default if you prefer:
+
+```json
+{
+  "defaultAgent": "gemini"
+}
+```
+
+## When Extensions Are Ready
+
+Once Gemini extensions are implemented, you'll be able to use:
+
+```bash
+# Install extensions (future)
+relentless init --agent gemini
+# or
+gemini extensions install https://github.com/ArvorCo/Relentless
+
+# Use extensions (future)
+@constitution create for TypeScript project
+@specify Add user authentication
+@plan using React and PostgreSQL
+@tasks generate user stories
+@checklist create validation
+```
+
+## Best Practices for Gemini Users
+
+1. **Reference skill files in prompts:**
+   ```bash
+   gemini "Read .claude/skills/tasks/SKILL.md and follow its instructions to create tasks.md"
+   ```
+
+2. **Be explicit about file locations:**
+   ```bash
+   gemini "Create relentless/features/001-user-auth/tasks.md with these exact user stories..."
+   ```
+
+3. **Use the orchestrator for implementation:**
+   ```bash
+   relentless run --agent gemini --feature 001-user-auth --tui
+   ```
+
+4. **Leverage Gemini's strengths:**
+   - Good at reading and understanding structured instructions
+   - Can parse and follow complex formats
+   - Works well with file system operations
+
+## Contributing
+
+Help us add native Gemini extensions support:
+
+1. **Research:** Check Gemini's extension format and requirements
+2. **Convert:** Create script to convert `.claude/skills/` to `.gemini/extensions/`
+3. **Test:** Validate extensions work with Gemini CLI
+4. **Document:** Add usage examples and guides
+5. **Submit:** Create PR with working implementation
+
+## Resources
+
+- [Gemini CLI Documentation](https://github.com/google-gemini/gemini-cli)
+- [Relentless Skills Format](./.claude/skills/)
+- [Agent Adapter](./src/agents/gemini.ts)
+- [Manual Workflow Example](#quick-start-current)

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Arvor
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.