@jaguilar87/gaia-ops 2.1.0 → 2.2.1

package/CHANGELOG.md

@@ -7,6 +7,133 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ---
 
+## [2.2.1] - 2025-11-10
+
+### Fixed - Documentation Consistency
+- **README.md & README.en.md:**
+  - Updated version numbers from 2.1.0 → 2.2.0
+  - Corrected package structure (hooks/, templates/, commands/)
+  - Fixed hooks/ listing: now shows actual Python files (pre_tool_use.py, post_tool_use.py, etc.) instead of non-existent pre-commit
+  - Fixed templates/ listing: removed non-existent code-examples/, listed actual files (CLAUDE.template.md, settings.template.json)
+  - Added context-contracts.gcp.json and context-contracts.aws.json to config/ section
+  - Removed CLAUDE.md and AGENTS.md from package root (only templates exist)
+  - Added speckit/ directory to structure
+
+- **config/AGENTS.md:**
+  - Updated all references: `.claude/docs/` → `.claude/config/`
+  - Fixed quick links and support documentation paths
+
+- **config/agent-catalog.md:**
+  - Updated all 5 context contract references: `.claude/docs/` → `.claude/config/`
+
+- **index.js:**
+  - Deprecated `getDocPath()` function with console warning
+  - Function now redirects to `config/` directory instead of non-existent `docs/`
+  - Added JSDoc @deprecated annotation
+
+- **README.en.md (Documentation section):**
+  - Removed broken reference to `./CLAUDE.md` (file not in package)
+  - Fixed all documentation links: `./docs/` → `./config/`
+  - Updated to match actual config/ directory structure
+
+- **speckit/README.en.md:**
+  - Removed 3 non-existent commands: speckit.clarify, speckit.analyze-plan, speckit.constitution
+  - Updated command count: 9 → 7 actual commands
+  - Removed references to non-existent tasks-richer.py tool
+  - Removed entire sections for non-existent templates (data-model-template.md, contracts-template.md)
+  - Updated tool files list with actual tools (task_manager.py, clarify_engine.py, context_provider.py)
+  - Fixed all code examples to use only existing commands
+
+- **tools/context_provider.py:**
+  - Added auto-detection for project-context.json location
+  - Honors GAIA_CONTEXT_PATH environment variable
+  - Falls back through common locations (.claude/project-context.json, .claude/project-context/project-context.json)
+  - Fixes agent routing failures when project-context.json is in non-legacy location
+
+- **package.json:**
+  - Fixed `npm test` script (was calling non-existent pytest tests)
+  - Now echoes informative message about fixture availability
+
+- **Agent Branding Unification:**
+  - Renamed `agents/claude-architect.md` → `agents/gaia.md` (aligns with gaia-ops package name)
+  - Renamed `commands/gaina.md` → `commands/gaia.md` (unified as `/gaia` command)
+  - Updated all references in README.md, README.en.md, and agents/gaia.md
+  - Complete branding consistency: package name, agent name, and command name all use "gaia"
+
+### Benefits
+- ✅ **Accurate documentation:** All paths and structures match actual package contents
+- ✅ **No broken links:** References point to existing files
+- ✅ **Clear API:** Deprecated functions clearly marked
+- ✅ **User trust:** Documentation matches reality
+- ✅ **npm test passes:** No false failures
+
+---
+
+## [2.2.0] - 2025-11-10
+
+### Added - Unified Settings Template & Auto-Installation
+- **NEW:** Created unified `templates/settings.template.json` (214 lines)
+  - Merged functionality from `settings.json` + `settings.local.json`
+  - Includes all hooks (PreToolUse, PostToolUse, SubagentStop)
+  - Complete permissions (75+ allow, 9 deny, 27 ask entries)
+  - Full security tier definitions (T0-T3)
+  - Environment configuration
+
+- **Auto-Installation:** `gaia-init.js` now automatically generates `.claude/settings.json`
+  - Added `generateSettingsJson()` function
+  - Integrated into installation workflow (Step 6.5)
+  - Projects get complete settings from day 1
+
+### Removed - Dead Code Elimination
+- **CLAUDE.md** from package root (only template exists now)
+- **templates/code-examples/** (321 lines - never imported or executed)
+  - `commit_validation.py`
+  - `clarification_workflow.py`
+  - `approval_gate_workflow.py`
+- **templates/project-context.template.json** (126 lines - unused, installer generates programmatically)
+- **templates/project-context.template.aws.json** (128 lines - never used)
+- **package.json:** Removed `CLAUDE.md` from files array
+
+### Changed - Package Consistency
+- **templates/CLAUDE.template.md:**
+  - Updated all references: `.claude/docs/` → `.claude/config/`
+  - Updated package name: `@aaxis/claude-agents` → `@jaguilar87/gaia-ops`
+  - Removed code-examples reference (no longer exists)
+
+- **README.en.md:**
+  - Updated API examples to use `@jaguilar87/gaia-ops`
+  - Changed `getDocPath()` → `getConfigPath()` (correct function)
+
+- **index.js:**
+  - Updated header and JSDoc comments with new package name
+  - Updated example usage
+
+- **agents/gaia.md:**
+  - Updated system paths to reflect gaia-ops package structure
+  - Clarified symlink architecture and layout
+
+### Improved - Package Quality
+- **Reduced template bloat by 57%:** 882 lines → 378 lines (504 lines removed)
+- **Single source of truth:** One settings template instead of scattered config
+- **Cleaner architecture:** Only actual templates remain in `templates/`
+- **Better defaults:** Projects start with complete, production-ready settings
+
+### Benefits
+- ✅ **Unified configuration:** Everything in one settings.json file
+- ✅ **Automatic setup:** No manual settings configuration needed
+- ✅ **Smaller package:** 57% reduction in template code
+- ✅ **Flexibility maintained:** Users can still create `settings.local.json` for overrides
+- ✅ **Package consistency:** All references use correct package name
+
+### Final Template Structure
+```
+templates/
+├── CLAUDE.template.md         (164 lines) - Orchestrator instructions
+└── settings.template.json     (214 lines) - Complete Claude Code settings
+```
+
+---
+
 ## [2.1.0] - 2025-11-10
 
 ### Added - Provider-Specific Context Contracts

package/README.en.md

@@ -14,8 +14,9 @@ Multi-agent orchestration system for Claude Code - DevOps automation toolkit.
 
 ### Features
 
-- **6 specialist agents** (terraform-architect, gitops-operator, gcp-troubleshooter, aws-troubleshooter, devops-developer, claude-architect)
-- **3 meta-agents** (Explore, Plan, claude-architect)
+- **Multi-cloud support** - Works with GCP, AWS, and ready for Azure
+- **6 specialist agents** (terraform-architect, gitops-operator, gcp-troubleshooter, aws-troubleshooter, devops-developer, Gaia)
+- **3 meta-agents** (Explore, Plan, Gaia)
 - **Clarification engine** for ambiguity detection
 - **Approval gates** for T3 operations (terraform apply, kubectl apply, etc.)
 - **Git commit validation** with Conventional Commits
@@ -90,7 +91,7 @@ node_modules/@jaguilar87/gaia-ops/
 │   ├── gcp-troubleshooter.md
 │   ├── aws-troubleshooter.md
 │   ├── devops-developer.md
-│   └── claude-architect.md
+│   └── gaia.md
 ├── tools/               # Orchestration tools
 │   ├── context_provider.py
 │   ├── agent_router.py
@@ -98,28 +99,35 @@ node_modules/@jaguilar87/gaia-ops/
 │   ├── approval_gate.py
 │   ├── commit_validator.py
 │   └── task_manager.py
-├── hooks/               # Git hooks
-│   └── pre-commit
+├── hooks/               # Claude Code hooks (Python)
+│   ├── pre_tool_use.py
+│   ├── post_tool_use.py
+│   ├── subagent_stop.py
+│   ├── session_start.py
+│   └── pre_kubectl_security.py
 ├── commands/            # Slash commands
-│   ├── architect.md
-│   └── speckit.*.md
+│   ├── gaia.md
+│   ├── save-session.md
+│   ├── restore-session.md
+│   ├── session-status.md
+│   └── speckit.*.md (7 commands)
 ├── config/              # Configuration & documentation
 │   ├── AGENTS.md
 │   ├── orchestration-workflow.md
 │   ├── git-standards.md
 │   ├── context-contracts.md
+│   ├── context-contracts.gcp.json
+│   ├── context-contracts.aws.json
 │   ├── agent-catalog.md
 │   └── git_standards.json
-├── templates/           # Code templates
+├── templates/           # Installation templates
 │   ├── CLAUDE.template.md
-│   └── code-examples/
-│       ├── commit_validation.py
-│       ├── clarification_workflow.py
-│       └── approval_gate_workflow.py
-├── config/              # Configuration
-│   └── git_standards.json
-├── CLAUDE.md            # Core orchestrator instructions
-├── AGENTS.md            # System overview
+│   └── settings.template.json
+├── speckit/             # Spec-Kit methodology
+│   ├── README.md
+│   ├── templates/
+│   ├── scripts/
+│   └── governance.md
 ├── CHANGELOG.md         # Version history
 ├── package.json
 └── index.js             # Helper functions
@@ -160,17 +168,17 @@ If you need to programmatically access paths in the package:
 import {
   getAgentPath,
   getToolPath,
-  getDocPath
-} from '@aaxis/claude-agents';
+  getConfigPath
+} from '@jaguilar87/gaia-ops';
 
 const agentPath = getAgentPath('gitops-operator');
-// → /path/to/node_modules/@aaxis/claude-agents/agents/gitops-operator.md
+// → /path/to/node_modules/@jaguilar87/gaia-ops/agents/gitops-operator.md
 
 const toolPath = getToolPath('context_provider.py');
-// → /path/to/node_modules/@aaxis/claude-agents/tools/context_provider.py
+// → /path/to/node_modules/@jaguilar87/gaia-ops/tools/context_provider.py
 
-const docPath = getDocPath('orchestration-workflow.md');
-// → /path/to/node_modules/@aaxis/claude-agents/docs/orchestration-workflow.md
+const configPath = getConfigPath('orchestration-workflow.md');
+// → /path/to/node_modules/@jaguilar87/gaia-ops/config/orchestration-workflow.md
 ```
 
 ## Versioning
@@ -181,18 +189,17 @@ This package follows [Semantic Versioning](https://semver.org/):
 - **MINOR:** New features, agents, or improvements
 - **PATCH:** Bug fixes, clarifications, typos
 
-Current version: **2.1.0**
+Current version: **2.2.0**
 
 See [CHANGELOG.md](./CHANGELOG.md) for version history.
 
 ## Documentation
 
-- **Core Instructions:** [CLAUDE.md](./CLAUDE.md) (154 lines)
-- **System Overview:** [AGENTS.md](./AGENTS.md) (95 lines)
-- **Orchestration Workflow:** [docs/orchestration-workflow.md](./docs/orchestration-workflow.md) (735 lines)
-- **Git Standards:** [docs/git-standards.md](./docs/git-standards.md) (682 lines)
-- **Context Contracts:** [docs/context-contracts.md](./docs/context-contracts.md) (673 lines)
-- **Agent Catalog:** [docs/agent-catalog.md](./docs/agent-catalog.md) (603 lines)
+- **System Overview:** [config/AGENTS.md](./config/AGENTS.md) (95 lines)
+- **Orchestration Workflow:** [config/orchestration-workflow.md](./config/orchestration-workflow.md) (735 lines)
+- **Git Standards:** [config/git-standards.md](./config/git-standards.md) (682 lines)
+- **Context Contracts:** [config/context-contracts.md](./config/context-contracts.md) (673 lines)
+- **Agent Catalog:** [config/agent-catalog.md](./config/agent-catalog.md) (603 lines)
 
 ## Requirements
 

package/README.md

@@ -14,8 +14,9 @@ Sistema de orquestación multi-agente para Claude Code - Toolkit de automatizaci
 
 ### Características
 
-- **6 agentes especialistas** (terraform-architect, gitops-operator, gcp-troubleshooter, aws-troubleshooter, devops-developer, claude-architect)
-- **3 meta-agentes** (Explore, Plan, claude-architect)
+- **Soporte multi-cloud** - Funciona con GCP, AWS, y listo para Azure
+- **6 agentes especialistas** (terraform-architect, gitops-operator, gcp-troubleshooter, aws-troubleshooter, devops-developer, Gaia)
+- **3 meta-agentes** (Explore, Plan, Gaia)
 - **Motor de clarificación** para detección de ambigüedades
 - **Puertas de aprobación** para operaciones T3 (terraform apply, kubectl apply, etc.)
 - **Validación de commits Git** con Conventional Commits
@@ -90,7 +91,7 @@ node_modules/@jaguilar87/gaia-ops/
 │   ├── gcp-troubleshooter.md
 │   ├── aws-troubleshooter.md
 │   ├── devops-developer.md
-│   └── claude-architect.md
+│   └── gaia.md
 ├── tools/               # Herramientas de orquestación
 │   ├── context_provider.py
 │   ├── agent_router.py
@@ -98,28 +99,35 @@ node_modules/@jaguilar87/gaia-ops/
 │   ├── approval_gate.py
 │   ├── commit_validator.py
 │   └── task_manager.py
-├── hooks/               # Git hooks
-│   └── pre-commit
+├── hooks/               # Hooks de Claude Code (Python)
+│   ├── pre_tool_use.py
+│   ├── post_tool_use.py
+│   ├── subagent_stop.py
+│   ├── session_start.py
+│   └── pre_kubectl_security.py
 ├── commands/            # Comandos slash
-│   ├── architect.md
-│   └── speckit.*.md
+│   ├── gaia.md
+│   ├── save-session.md
+│   ├── restore-session.md
+│   ├── session-status.md
+│   └── speckit.*.md (7 comandos)
 ├── config/              # Configuración y documentación
 │   ├── AGENTS.md
 │   ├── orchestration-workflow.md
 │   ├── git-standards.md
 │   ├── context-contracts.md
+│   ├── context-contracts.gcp.json
+│   ├── context-contracts.aws.json
 │   ├── agent-catalog.md
 │   └── git_standards.json
-├── templates/           # Plantillas de código
+├── templates/           # Templates de instalación
 │   ├── CLAUDE.template.md
-│   └── code-examples/
-│       ├── commit_validation.py
-│       ├── clarification_workflow.py
-│       └── approval_gate_workflow.py
-├── config/              # Configuración
-│   └── git_standards.json
-├── CLAUDE.md            # Instrucciones del orquestador principal
-├── AGENTS.md            # Vista general del sistema
+│   └── settings.template.json
+├── speckit/             # Metodología Spec-Kit
+│   ├── README.md
+│   ├── templates/
+│   ├── scripts/
+│   └── governance.md
 ├── CHANGELOG.md         # Historial de versiones
 ├── package.json
 └── index.js             # Funciones auxiliares
@@ -181,7 +189,7 @@ Este paquete sigue [Versionamiento Semántico](https://semver.org/):
 - **MINOR:** Nuevas características, agentes o mejoras
 - **PATCH:** Correcciones de bugs, clarificaciones, errores tipográficos
 
-Versión actual: **1.0.3**
+Versión actual: **2.2.0**
 
 Ver [CHANGELOG.md](./CHANGELOG.md) para el historial de versiones.
 

package/agents/{claude-architect.md → gaia.md}

@@ -1,11 +1,11 @@
 ---
-name: claude-architect
-description: A meta-agent specialized in analyzing, diagnosing, and optimizing the intelligent agent orchestration system itself. It understands the system architecture, analyzes logs/metrics, researches best practices, and proposes improvements.
+name: gaia
+description: Gaia is the meta-agent for the gaia-ops ecosystem—its purpose is to understand, document, and continuously optimize every component of the orchestrator, agents, commands, templates, sample stacks, and documentation that ship in this package and the consuming projects that symlink to it.
 tools: Read, Glob, Grep, Bash, Task, WebSearch, Python
 model: inherit
 ---
 
-You are a senior system architect and AI agent systems specialist. Your unique purpose is to **analyze and optimize the intelligent agent orchestration system itself** - acting as a meta-layer that understands how the orchestrator, agents, router, context provider, and all system components work together.
+You are Gaia, the senior system architect and AI agent systems specialist for the gaia-ops stack. Your unique purpose is to **analyze and optimize Gaia’s intelligent agent orchestration system end-to-end**—acting as a meta-layer that understands how the orchestrator, agents, commands, router, context provider, Spec-Kit assets, and all system components work together across repositories (package root, downstream project `.claude/`, ops symlinks).
 
 ## ⚡ QUICK START - Read This First
 
@@ -15,16 +15,16 @@ You are a senior system architect and AI agent systems specialist. Your unique p
 2. **Locate & Read:** You know where EVERYTHING lives. Read only what you need for THIS request.
 3. **Analyze & Respond:** Provide comprehensive answer with evidence, examples, and actionable recommendations.
 
-**Where Everything Lives (You Know This By Heart):**
-- 🏗️ System: `/home/jaguilar/aaxis/rnd/repositories/.claude/`
-- 📋 Orchestrator: `CLAUDE.md` (workflow logic)
-- 🤖 Agents: `.claude/agents/` (5 specialists + you)
-- 🛠️ Tools: `.claude/tools/` (routing, context, validation)
-- 📊 Logs: `.claude/logs/` (JSONL audit trail)
-- ✅ Tests: `.claude/tests/` (55+ tests)
-- 🎯 Spec-Kit: `.claude/commands/speckit.*` (7 commands)
-- 💾 Sessions: `.claude/session/` (active + bundles)
-- 🔗 Multi-repo: `ops/` (symlinks: claude-rnd, claude-vtr)
+**Where Everything Lives (Package + Symlink Layout):**
+- 🏗️ Package root: `/home/jaguilar/aaxis/rnd/repositories/gaia-ops/` → mirrors `node_modules/@jaguilar87/gaia-ops/` when installed
+- 📋 Orchestrator: `CLAUDE.md` at the package root (templated into consuming repos)
+- 🤖 Agents: `agents/*.md` (6 specialists: terraform-architect, gitops-operator, gcp-troubleshooter, aws-troubleshooter, devops-developer, gaia)
+- 🛠️ Tools: `tools/` (context_provider.py, agent_router.py, clarify_engine.py, approval_gate.py, commit_validator.py, task_manager.py)
+- 📚 Config docs: `config/` (AGENTS, orchestration-workflow, git-standards, context-contracts, agent-catalog)
+- 🗂️ Commands: `commands/*.md` (`/gaia`, `/save-session`, `/session-status`, `/speckit.*`)
+- 🎯 Spec-Kit assets: `speckit/README*.md`, `speckit/templates/`, `speckit/scripts/`, `speckit/governance.md`, `speckit/decisions/`
+- 🔧 Reference stacks: `terraform/`, `gitops/`, `app-services/` illustrate how agents interact with user IaC/App repos
+- 💾 Project data: consuming repos host `.claude/project-context.json`, `.claude/logs/`, `.claude/tests/`, while `ops/` carries shared symlink helpers
 
 **Your Superpowers:**
 - ✅ You understand the ENTIRE system (no one else does)
@@ -65,43 +65,51 @@ You have intrinsic knowledge of the system's structure. You know EXACTLY where t
 ### Core System Files (Always Available)
 
 ```
-Agent System Structure:
-├── CLAUDE.md                           # Master orchestrator logic (715 lines)
-├── .claude/
-│   ├── project-context.json           # Project SSOT (varies by project)
-│   ├── settings.json                  # System configuration
-│   ├── agents/                        # 5 specialized agents
-│   │   ├── gitops-operator.md         (340 lines)
-│   │   ├── terraform-architect.md     (270 lines)
-│   │   ├── gcp-troubleshooter.md      (305 lines)
-│   │   ├── aws-troubleshooter.md      (289 lines)
-│   │   └── devops-developer.md        (89 lines)
-│   ├── tools/                         # System intelligence
-│   │   ├── agent_router.py            # Semantic routing (92.7% accuracy target)
-│   │   ├── context_provider.py        # Deterministic context generation
-│   │   ├── context_section_reader.py  # Selective context loading
-│   │   ├── semantic_matcher.py        # Fallback routing
-│   │   ├── agent_invoker_helper.py    # Agent invocation utilities
-│   │   ├── tasks-richer.py            # Task enrichment
-│   │   └── generate_embeddings.py     # Embedding generation
-│   ├── hooks/                         # Security & audit layer
-│   │   ├── pre_tool_use.py           # Pre-execution validation
-│   │   ├── post_tool_use.py          # Post-execution audit
-│   │   └── subagent_stop.py          # Agent completion capture
-│   ├── commands/                      # 13 slash commands
-│   ├── session/                       # Session management
-│   │   ├── active/context.json       # Live session state
-│   │   ├── bundles/                  # Historical snapshots
-│   │   └── scripts/                  # Session tools
-│   ├── tests/                         # Test suite (55+ tests)
-│   │   ├── test_semantic_routing.py  # Routing accuracy tests
-│   │   ├── test_all_functionality.py # Core system tests
-│   │   └── test_ssot_policies.py     # SSOT validation
-│   ├── logs/                          # Audit trail (JSONL format)
-│   └── schemas/                       # JSON schemas
-└── improvement-ideas.md               # System improvement backlog
+gaia-ops/  (mirrors node_modules/@jaguilar87/gaia-ops/ and symlinks into project .claude/)
+├── CLAUDE.md                       # Master orchestrator logic + workflow (≈150 lines here, expanded in config/)
+├── agents/                         # 6 specialized agent prompts
+│   ├── terraform-architect.md
+│   ├── gitops-operator.md
+│   ├── gcp-troubleshooter.md
+│   ├── aws-troubleshooter.md
+│   ├── devops-developer.md
+│   └── gaia.md (this file)
+├── tools/                          # System intelligence + automation
+│   ├── context_provider.py         # Deterministic context generation
+│   ├── agent_router.py             # Semantic routing (92.7% target accuracy)
+│   ├── clarify_engine.py           # Ambiguity detection
+│   ├── approval_gate.py            # Tiered approval logic
+│   ├── commit_validator.py         # Conventional commits enforcement
+│   └── task_manager.py             # Large-plan chunking
+├── hooks/                          # Git + security hooks
+├── commands/                       # User-facing slash commands (architect/save-session/speckit.*)
+├── config/                         # Documentation bundle (AGENTS/orchestration-workflow/git-standards/context-contracts/agent-catalog)
+├── speckit/                        # Spec-Kit 2.0 framework (README*.md, governance.md, decisions/, templates/, scripts/)
+├── app-services/                   # Sample application services for reference
+├── gitops/                         # Reference GitOps manifests
+├── terraform/                      # Reference Terraform stacks
+├── templates/                      # CLAUDE + code templates
+├── tests/                          # Test suite (55+ cases referenced in README)
+├── CHANGELOG.md                    # Version history (Semantic Versioning)
+└── package.json / index.js         # NPM package metadata + helper exports
 ```
 
+When `npx @jaguilar87/gaia-ops init` (or `gaia-init` after a global install) runs in a consuming project it:
+1. Detects GitOps/Terraform/AppServices paths and installs Claude Code if needed.
+2. Creates `.claude/` and symlinks `agents/`, `tools/`, `hooks/`, `commands/`, `templates/`, and `config/` back to this package.
+3. Generates `CLAUDE.md`, links `AGENTS.md`, and seeds `.claude/project-context.json` (project SSOT). Project-owned items such as `.claude/logs/`, `.claude/tests/`, and session data remain local.
+
+### Installation & Project Layout (from README.md / README.en.md)
+- Quick start: `npx @jaguilar87/gaia-ops init` (or `npm install -g @jaguilar87/gaia-ops && gaia-init`) bootstraps everything; manual installs `npm install @jaguilar87/gaia-ops` + symlink commands from README.
+- Resulting structure: `your-project/.claude/{agents,tools,hooks,commands,templates,config}` → symlinked to this package under `node_modules/@jaguilar87/gaia-ops/`, while `logs/`, `tests/`, and `project-context.json` stay project-specific.
+- Reference directories `gitops/`, `terraform/`, and `app-services/` inside the package illustrate how specialized agents should reason about user IaC/App codebases.
+
+### Spec-Kit 2.0 Workflow Snapshot (from speckit/README*.md)
+- `speckit/` hosts bilingual docs, governance (`speckit/governance.md`), immutable ADRs (`speckit/decisions/ADR-*.md`), templates, and scripts backing `/speckit.*` commands.
+- Core flow: `/speckit.init` → `/speckit.specify` → `/speckit.plan` → `/speckit.tasks` → `/speckit.implement`, each auto-injecting project-context data, clarification, validation, and risk analysis (T2/T3 gates) directly into specs/plan/tasks artifacts.
+- Helper commands: `/speckit.add-task` adds enriched tasks mid-implementation, `/speckit.analyze-task` deep-dives high-risk tasks, and `/save-session` captures context bundles for portability.
+- Spec-Kit 2.0 removes standalone enrichers (`tasks-richer.py`), performs inline validation, and relies on `.claude/project-context.json` for deterministic context so the architect agent can reason about idea → spec → plan → tasks → implementation continuity.
+
 ### Key System Metrics (What to Track)
 
 - **Routing Accuracy:** Target 92.7% (from tests)
@@ -818,4 +826,3 @@ python3 -m pytest .claude/tests/ --cov=.claude/tools --cov-report=term
 ---
 
 **Remember:** You are not just analyzing files - you are understanding a living, evolving system. Your insights drive its continuous improvement.
-

package/bin/gaia-init.js

@@ -603,6 +603,29 @@ async function createClaudeDirectory() {
   }
 }
 
+/**
+ * Generate settings.json from template
+ */
+async function generateSettingsJson() {
+  const spinner = ora('Generating .claude/settings.json...').start();
+
+  try {
+    const settingsPath = join(CWD, '.claude', 'settings.json');
+    const templatePath = getTemplatePath('settings.template.json');
+
+    // Read template
+    const template = await fs.readFile(templatePath, 'utf-8');
+
+    // Write settings.json (no placeholders to replace)
+    await fs.writeFile(settingsPath, template, 'utf-8');
+
+    spinner.succeed('.claude/settings.json generated');
+  } catch (error) {
+    spinner.fail('Failed to generate settings.json');
+    throw error;
+  }
+}
+
 /**
  * Generate CLAUDE.md from template
  */
@@ -1000,6 +1023,9 @@ async function main() {
     // Step 6: Create .claude/ directory with symlinks
     await createClaudeDirectory();
 
+    // Step 6.5: Generate settings.json
+    await generateSettingsJson();
+
     // Step 7: Generate CLAUDE.md
     await generateClaudeMd(config);
 

package/commands/{architect.md → gaia.md}

@@ -1,13 +1,13 @@
 ---
-description: Invoke claude-architect to analyze system architecture, diagnose issues, or propose improvements
+description: Invoke Gaia to analyze system architecture, diagnose issues, or propose improvements
 scope: project
 ---
 
-You are invoking the **claude-architect** meta-agent to analyze the agent orchestration system itself.
+You are invoking **Gaia**, the gaia-ops meta-agent that analyzes the agent orchestration system itself.
 
 ## Context: System Expert Agent
 
-claude-architect is a specialized agent that understands:
+Gaia is a specialized agent that understands:
 - The entire agent system architecture (.claude/, CLAUDE.md, agents, tools, hooks)
 - Spec-Kit workflows (specify, plan, tasks, implement, add-task, analyze-task)
 - Session management (save, restore, bundles)
@@ -18,7 +18,7 @@ claude-architect is a specialized agent that understands:
 
 ## Your Task
 
-Invoke the claude-architect agent with the user's request. The agent will proactively read system files, analyze logs, research best practices, and provide comprehensive analysis with actionable recommendations.
+Invoke Gaia with the user's request. The agent will proactively read system files, analyze logs, research best practices, and provide comprehensive analysis with actionable recommendations tied to gaia-ops.
 
 **IMPORTANT:** Pass the user's exact question/request to the agent. Examples:
 - "Analiza este log y dime qué pasó"
@@ -29,11 +29,11 @@ Invoke the claude-architect agent with the user's request. The agent will proact
 
 ## Invocation
 
-Use the Task tool to invoke claude-architect with this structure:
+Use the Task tool to invoke Gaia with this structure:
 
 ```
 Task(
-    subagent_type="claude-architect",
+    subagent_type="gaia",
     description="[Brief 3-5 word description]",
     prompt="""
 ## System Context (Auto-provided)

package/config/AGENTS.md

@@ -8,9 +8,9 @@ See `CLAUDE.md` for complete orchestrator instructions.
 
 **Quick links:**
 - Core principles: `CLAUDE.md` (lines 18-36)
-- Workflow: `.claude/docs/orchestration-workflow.md`
-- Git standards: `.claude/docs/git-standards.md`
-- Agent catalog: `.claude/docs/agent-catalog.md`
+- Workflow: `.claude/config/orchestration-workflow.md`
+- Git standards: `.claude/config/git-standards.md`
+- Agent catalog: `.claude/config/agent-catalog.md`
 
 ---
 
@@ -71,7 +71,7 @@ Claude Code (Orchestrator)
     ├── gcp-troubleshooter (GCP diagnostics)
     ├── aws-troubleshooter (AWS diagnostics)
     ├── devops-developer (Application build/test)
-    ├── claude-architect (System optimization)
+    ├── Gaia (System optimization)
     ├── Explore (Codebase exploration)
     └── Plan (Implementation planning)
 ```
@@ -153,7 +153,7 @@ See `CLAUDE.md` and `.claude/CHANGELOG.md` for contribution guidelines.
 
 - **Issues:** Create issue in GitHub repository
 - **Questions:** Contact Jorge Aguilar (jaguilar@aaxis.com)
-- **Documentation:** See `.claude/docs/*.md`
+- **Documentation:** See `.claude/config/*.md`
 
 ---