chati-dev 3.3.2 → 4.0.1

package/README.md

@@ -266,7 +266,7 @@ Three systems operate transparently behind the pipeline:
 
 | System | What it does |
 |--------|-------------|
-| **Context Engine (PRISM)** | Injects the right context at the right time. 5 layers of context (from system-wide rules down to specific task details). Tracks how much context space remains and adapts automatically. |
+| **Context Engine (PRISM)** | Injects the right context at the right time. 6 layers of context (from constitution rules down to keyword-matched domain rules). Uses Progressive Reinforcement: as context degrades, injection budgets INCREASE (1.5% to 5.0% of provider context window). Provider-aware for Claude (200K), Gemini (1M), and Codex (128K). |
 | **Memory System (RECALL)** | Remembers decisions, gotchas, and lessons across sessions. Organized into 4 sectors: what happened (episodic), what we know (semantic), how we do things (procedural), and what we learned (reflective). |
 | **Decision Engine (COMPASS)** | Before creating something new, checks if a similar component already exists. Decides whether to reuse, adapt, or create from scratch. Keeps a registry of all project entities. |
 
@@ -390,8 +390,8 @@ your-project/
 │   │   ├── quality/              # QA-Planning, QA-Implementation
 │   │   ├── build/                # Dev
 │   │   └── deploy/               # DevOps
-│   ├── workflows/                # 6 workflow blueprints
-│   ├── templates/                # 6 artifact templates
+│   ├── workflows/                # 7 workflow blueprints
+│   ├── templates/                # 10 artifact templates
 │   ├── schemas/                  # 5 JSON schemas
 │   ├── intelligence/             # PRISM, RECALL, COMPASS specs
 │   ├── domains/                  # Per-agent and per-workflow configs
@@ -454,5 +454,5 @@ For security concerns, see our [Security Policy](.github/SECURITY.md).
 
 <p align="center">
   <sub>Built with structure, validated by agents, governed by constitution.</sub><br>
-  <sub>Chati.dev &copy; 2026</sub>
+  <sub>Chati.dev v4.0 "Pegasus" &copy; 2026</sub>
 </p>

package/framework/agents/plan/ux.md

@@ -43,7 +43,47 @@ Define the user experience: information architecture, user flows, interaction pa
 
 ---
 
-## Execution: 5 Phases
+## Execution: 6 Phases
+
+### Phase 0: Brand Identity (Brandbook)
+```
+PREREQUISITE: Brief handoff + Architecture handoff
+
+Before defining any design tokens or visual patterns, establish the brand identity:
+
+1. Define Brand Voice
+   - Personality traits (3-5 adjectives)
+   - Tone spectrum (formal <-> casual positioning)
+   - Writing principles and anti-patterns
+
+2. Define Visual Language
+   - Mood descriptor and visual metaphor
+   - Density and contrast philosophy
+   - Texture strategy (flat, layered, glassmorphic, etc.)
+
+3. Typography Rationale
+   - Display + body font pairing with WHY
+   - Hierarchy strategy
+   - Pairing logic explanation
+
+4. Color Philosophy
+   - Primary color rationale (emotional, functional)
+   - Palette derivation logic
+   - Accessibility considerations
+
+5. Spacing Rhythm
+   - Base unit selection + rationale
+   - Rhythm personality (generous, dense, modular)
+
+6. Iconography Style
+   - Style direction (outline, solid, duo-tone)
+   - Preferred library alignment with brand personality
+
+Output: chati.dev/artifacts/4-UX/brandbook.md
+
+RULE: Every subsequent design decision must trace back to the brandbook.
+RULE: Brandbook establishes the WHY — Design System (Phase 4) establishes the WHAT.
+```
 
 ### Phase 1: User Flow Mapping
 ```

package/framework/config.yaml

@@ -1,8 +1,8 @@
 # chati.dev Configuration
-version: "3.3.2"
+version: "4.0.1"
 installed_at: "2026-02-07T10:00:00Z"
-updated_at: "2026-03-03T00:00:00Z"
-installer_version: "3.3.2"
+updated_at: "2026-03-18T00:00:00Z"
+installer_version: "4.0.1"
 project_type: greenfield
 language: en
 ides: [claude-code]
@@ -28,6 +28,21 @@ providers:
 #     provider: gemini
 #     model: pro
 
+# Feature flags — toggle new capabilities (v4.0 Pegasus)
+features:
+  hybrid_budget: true
+  anti_dash: true
+  rate_limiter_integration: true
+  l5_keywords: true
+  prompt_size_guard: true
+  ids_decision_engine: true
+  surface_criteria: true
+  parallel_fallback: true
+  tool_mesh: true
+  tech_presets: true
+  doctor_autofix: true
+  brandbook: true
+
 # Telemetry — opt-in anonymous usage tracking
 telemetry:
   enabled: true

package/framework/constitution.md

@@ -32,13 +32,13 @@ Chati.dev is a planning-first AI-assisted orchestration system that coordinates
 
 ### Exclusion List
 
-The following items from source projects are explicitly excluded from Chati.dev:
-- CI/CD pipelines, release scripts, publish scripts from source projects
+The following items are explicitly out of scope for Chati.dev:
+- CI/CD pipelines, release scripts, publish scripts (contributor tooling)
 - Husky/lint-staged configs (contributor tooling)
 - npm package configs (.npmignore, .npmrc)
 - Dashboard web UI (separate product)
 - Monitor Server (separate product)
-- Source project installer packages
+- Standalone installer packages (CLI handles installation)
 - execution-n8n skill (removed — code-only focus)
 - Generic tasks redundant with agent self-validation protocols
 - Duplicate checklists already covered by agent self-validation
@@ -115,6 +115,12 @@ Every agent in Chati.dev:
 4. Error messages are constructive: what failed, why, and how to fix
 5. Status updates are concise and actionable
 6. Technical jargon is adapted to detected user level (vibecoder vs power user)
+7. Generated content MUST NOT contain AI-detectable writing patterns:
+   - No em dashes. Use commas, semicolons, colons, or periods instead
+   - Avoid formulaic phrases: "Delve into", "Dive into", "It's important to note that"
+   - Avoid stacking transitions: "Furthermore", "Moreover", "Additionally" in sequence
+   - Avoid over-hedging: "It could potentially perhaps maybe..."
+   - Write naturally. If a sentence reads like AI wrote it, rewrite it
 
 **Enforcement: GUIDE** — Violations trigger guidance, not blocking.
 
@@ -270,7 +276,8 @@ The pipeline operates in three execution modes that control agent permissions. M
 1. The orchestrator SHALL calculate the context bracket (FRESH, MODERATE, DEPLETED, CRITICAL) before every agent interaction.
 
 2. Context injection layers SHALL be reduced according to bracket level:
-   - FRESH/MODERATE: All 5 layers (L0-L4)
+   - FRESH: All 6 layers (L0-L5)
+   - MODERATE: L0 + L1 + L2 + L3 + L5 (skip L4 task detail)
    - DEPLETED: L0 (Constitution) + L1 (Mode) + L2 (Agent) only
    - CRITICAL: L0 (Constitution) + L1 (Mode) only
 
@@ -280,11 +287,12 @@ The pipeline operates in three execution modes that control agent permissions. M
 
 4. The Constitution (L0) and Mode governance (L1) are NON-NEGOTIABLE and SHALL be injected in ALL brackets, including CRITICAL.
 
-5. Token budgets per bracket:
-   - FRESH: 2500 tokens maximum
-   - MODERATE: 2000 tokens maximum
-   - DEPLETED: 1500 tokens maximum
-   - CRITICAL: 800 tokens maximum
+5. Token budgets per bracket (Progressive Reinforcement Model — as context degrades, reinforcement increases):
+   - Budgets are expressed as a percentage of the provider's context window
+   - FRESH: 1.5% of context window (Claude: 3000, Gemini: 15000, Codex: 1920)
+   - MODERATE: 2.5% of context window (Claude: 5000, Gemini: 25000, Codex: 3200)
+   - DEPLETED: 4.0% of context window (Claude: 8000, Gemini: 40000, Codex: 5120)
+   - CRITICAL: 5.0% of context window (Claude: 10000, Gemini: 50000, Codex: 6400)
 
 6. Autonomous spawn capability varies by IDE:
    - Full autonomy: Claude Code, AntiGravity, Gemini CLI
@@ -472,5 +480,5 @@ When multiple CLI providers are enabled, the system SHALL coordinate agent execu
 
 ---
 
-*Chati.dev Constitution v3.3.0 — 19 Articles + Preamble*
+*Chati.dev Constitution v4.0.0 — 19 Articles + Preamble*
 *All agents are bound by this Constitution. Violations are enforced per article.*

package/framework/context/governance.md

@@ -10,6 +10,8 @@ Extracted from `chati.dev/constitution.md` (19 Articles). Read the full constitu
 ## Context Bracket (Article XII)
 - 4 brackets: FRESH, MODERATE, DEPLETED, CRITICAL
 - CRITICAL = Constitution + Global layers only (no agent/task context)
+- Progressive Reinforcement: budget INCREASES as context degrades (1.5% → 2.5% → 4.0% → 5.0% of provider context window)
+- Provider-aware: budgets scale proportionally to context window size (Claude 200K, Gemini 1M, Codex 128K)
 - Handoff before bracket drops below 15%
 
 ## Memory (Article XIII)

package/framework/context/root.md

@@ -1,7 +1,7 @@
 # Chati.dev System Context
 
 ## Framework
-- **Version**: 3.3.0
+- **Version**: 4.0.0
 - **Agents**: 13 (12 specialized + orchestrator)
 - **Constitution**: 19 Articles + Preamble
 - **Quality**: 5 pipeline gates + 3-tier verdicts (APPROVED / NEEDS_REVISION / BLOCKED)

package/framework/data/entity-registry.yaml

@@ -3,9 +3,9 @@
 # and by the Health Check for system integrity validation.
 
 metadata:
-  version: "1.0.0"
-  last_updated: "2026-02-12T00:00:00Z"
-  entity_count: 50
+  version: "4.0.0"
+  last_updated: "2026-03-17T00:00:00Z"
+  entity_count: 55
   checksum_algorithm: sha256
 
 entities:
@@ -163,6 +163,38 @@ entities:
       dependencies: [brief]
       adaptability: 0.6
 
+    brandbook:
+      path: chati.dev/templates/brandbook-tmpl.yaml
+      type: template
+      purpose: "Brand identity foundation — voice, visual language, typography, color philosophy"
+      keywords: [brand, identity, voice, visual, typography, color]
+      dependencies: [ux]
+      adaptability: 0.8
+
+    design-tokens:
+      path: chati.dev/templates/design-token-tmpl.yaml
+      type: template
+      purpose: "Structured design token system — primitive, semantic, component, utility layers"
+      keywords: [design, tokens, colors, spacing, typography, components]
+      dependencies: [brandbook]
+      adaptability: 0.7
+
+    component-spec:
+      path: chati.dev/templates/component-spec-tmpl.yaml
+      type: template
+      purpose: "UI component specification — variants, states, anatomy, accessibility"
+      keywords: [component, specification, ui, variants, states, accessibility]
+      dependencies: [design-tokens]
+      adaptability: 0.7
+
+    icon-system:
+      path: chati.dev/templates/icon-system-tmpl.yaml
+      type: template
+      purpose: "Icon library selection, sizing scale, usage guidelines, and inventory"
+      keywords: [icons, library, sizing, accessibility, inventory]
+      dependencies: [brandbook]
+      adaptability: 0.8
+
   workflows:
     greenfield-fullstack:
       path: chati.dev/workflows/greenfield-fullstack.yaml
@@ -212,6 +244,14 @@ entities:
       dependencies: [orchestrator, brief, dev, qa-implementation, devops]
       adaptability: 0.4
 
+    standard-flow:
+      path: chati.dev/workflows/standard-flow.yaml
+      type: workflow
+      purpose: "Balanced pipeline for moderate tasks: features, integrations, brownfield enhancements. Uses 8 agents."
+      keywords: [standard, balanced, feature, integration, enhancement, moderate]
+      dependencies: [orchestrator, brief, detail, architect, tasks, qa-planning, dev, qa-implementation, devops]
+      adaptability: 0.4
+
   schemas:
     session:
       path: chati.dev/schemas/session.schema.json

package/framework/domains/constitution.yaml

@@ -29,7 +29,7 @@ rules:
     priority: critical
 
   - id: art-v
-    text: "Communication follows structured protocol: guided options (1,2,3), clear formatting."
+    text: "Communication follows structured protocol: guided options (1,2,3), clear formatting. No em dashes or AI-detectable writing patterns."
     priority: normal
 
   - id: art-vi

package/framework/domains/global.yaml

@@ -18,6 +18,10 @@ rules:
     text: "Interact with user in their preferred language. Artifacts stay in English."
     priority: normal
 
+  - id: anti-dash
+    text: "Never use em dashes in generated content. Avoid AI-detectable writing patterns (formulaic transitions, over-hedging, stacking connectors)."
+    priority: high
+
   # Execution Profile Rules (v3.0.0)
   execution-profiles:
     - explore-readonly: "In explore profile, all write/edit tool calls are blocked."
@@ -75,10 +79,10 @@ modes:
 
 brackets:
   FRESH:
-    behavior: "Full context injection. All layers active. Include detailed rules and examples."
+    behavior: "Moderate context injection. All layers active but model is warming up. Budget lower than MODERATE."
   MODERATE:
-    behavior: "Standard injection. Skip task detail layer (L4). Summarize long rules."
+    behavior: "Peak context injection. Maximum token budget. Model has context and needs maximum reinforcement."
   DEPLETED:
-    behavior: "Minimal injection. Only L0+L1+L2. Use rule IDs instead of full text."
+    behavior: "Reduced injection. Only L0+L1+L2. Use rule IDs instead of full text."
   CRITICAL:
     behavior: "Emergency. L0+L1 only. Trigger handoff advisory. Preserve essential state."

package/framework/domains/workflows/standard-flow.yaml

@@ -0,0 +1,33 @@
+# Standard Flow Workflow Domain — PRISM L3
+# Balanced pipeline for moderate tasks (features, integrations, brownfield enhancements)
+# Uses 8 agents — between Quick Flow (4) and Full Flow (12)
+
+steps:
+  - brief
+  - detail
+  - architect
+  - tasks
+  - qa-planning
+  - dev
+  - qa-implementation
+  - devops
+
+rules:
+  - id: sf-scope
+    text: "Standard Flow targets 2-5 requirements. Escalate to full pipeline if scope exceeds this range."
+    priority: critical
+  - id: sf-quality
+    text: "Both QA gates (planning and implementation) are mandatory. Score >= 95% required at each gate."
+    priority: critical
+  - id: sf-no-ux-skip
+    text: "UX agent is excluded. If design-heavy requirements surface, escalate to full pipeline."
+    priority: high
+  - id: sf-no-phases
+    text: "Phases agent is excluded. Tasks agent handles scope decomposition directly from PRD + Architecture."
+    priority: high
+  - id: sf-escalation-up
+    text: "If architect discovers multi-service complexity or QA finds UX gaps, offer escalation to full pipeline."
+    priority: high
+  - id: sf-escalation-down
+    text: "If detail finds only 1 simple requirement and architect confirms no changes needed, suggest quick flow."
+    priority: medium

package/framework/i18n/en.yaml

@@ -29,8 +29,8 @@ installer:
   created_overlays: "Created provider overlay directories"
   will_install: "Will install:"
   agents_count: "13 agent definitions (DISCOVER, PLAN, BUILD, DEPLOY phases)"
-  workflows_count: "6 workflow blueprints"
-  templates_count: "6 templates (PRD, Brownfield PRD, Architecture, Task, QA Gate, Quick Brief)"
+  workflows_count: "7 workflow blueprints"
+  templates_count: "10 templates (PRD, Brownfield PRD, Architecture, Task, QA Gate, Quick Brief, Brandbook, Design Tokens, Component Spec, Icon System)"
   constitution: "Constitution (19 Articles + Preamble)"
   session_mgmt: "Session management system"
   quality_gates: "Quality gates (3-tier verdicts)"
@@ -85,6 +85,15 @@ status:
   backlog_items: "Backlog Items"
   high_priority: "high priority"
 
+features:
+  brandbook_step: "Phase 0: Establishing brand identity (Brandbook)"
+  tech_preset_loaded: "Tech preset loaded: {preset}"
+  tech_preset_detected: "Detected tech stack: {stack}"
+  doctor_running: "Running Doctor diagnostics..."
+  doctor_healthy: "Doctor: All checks passed"
+  doctor_degraded: "Doctor: Issues found (see report)"
+  doctor_fix_applied: "Doctor: Applied {count} auto-fix(es)"
+
 errors:
   session_corrupted: "Session file appears corrupted. Attempting recovery..."
   handoff_missing: "Handoff not found. Using session.yaml + CLAUDE.md as fallback."

package/framework/i18n/es.yaml

@@ -29,8 +29,8 @@ installer:
   created_overlays: "Creados directorios de overlay de proveedores"
   will_install: "Se instalara:"
   agents_count: "13 definiciones de agentes (fases DISCOVER, PLAN, BUILD, DEPLOY)"
-  workflows_count: "6 blueprints de workflow"
-  templates_count: "6 templates (PRD, PRD Brownfield, Arquitectura, Task, QA Gate, Quick Brief)"
+  workflows_count: "7 blueprints de workflow"
+  templates_count: "10 templates (PRD, PRD Brownfield, Arquitectura, Task, QA Gate, Quick Brief, Brandbook, Design Tokens, Component Spec, Icon System)"
   constitution: "Constitucion (19 Articulos + Preambulo)"
   session_mgmt: "Sistema de gestion de sesion"
   quality_gates: "Quality gates (veredictos en 3 niveles)"
@@ -85,6 +85,15 @@ status:
   backlog_items: "Items del Backlog"
   high_priority: "alta prioridad"
 
+features:
+  brandbook_step: "Fase 0: Estableciendo identidad de marca (Brandbook)"
+  tech_preset_loaded: "Tech preset cargado: {preset}"
+  tech_preset_detected: "Stack tecnologico detectado: {stack}"
+  doctor_running: "Ejecutando diagnosticos del Doctor..."
+  doctor_healthy: "Doctor: Todos los checks pasaron"
+  doctor_degraded: "Doctor: Problemas encontrados (ver reporte)"
+  doctor_fix_applied: "Doctor: Aplicado(s) {count} auto-fix(es)"
+
 errors:
   session_corrupted: "Archivo de sesion parece corrupto. Intentando recuperacion..."
   handoff_missing: "Handoff no encontrado. Usando session.yaml + CLAUDE.md como respaldo."

package/framework/i18n/fr.yaml

@@ -29,8 +29,8 @@ installer:
   created_overlays: "Crees les repertoires d'overlay des fournisseurs"
   will_install: "Sera installe:"
   agents_count: "13 definitions d'agents (phases DISCOVER, PLAN, BUILD, DEPLOY)"
-  workflows_count: "6 blueprints de workflow"
-  templates_count: "6 templates (PRD, PRD Brownfield, Architecture, Task, QA Gate, Quick Brief)"
+  workflows_count: "7 blueprints de workflow"
+  templates_count: "10 templates (PRD, PRD Brownfield, Architecture, Task, QA Gate, Quick Brief, Brandbook, Design Tokens, Component Spec, Icon System)"
   constitution: "Constitution (19 Articles + Preambule)"
   session_mgmt: "Systeme de gestion de session"
   quality_gates: "Quality gates (verdicts en 3 niveaux)"
@@ -85,6 +85,15 @@ status:
   backlog_items: "Elements du Backlog"
   high_priority: "haute priorite"
 
+features:
+  brandbook_step: "Phase 0: Etablissement de l'identite de marque (Brandbook)"
+  tech_preset_loaded: "Tech preset charge: {preset}"
+  tech_preset_detected: "Stack technologique detecte: {stack}"
+  doctor_running: "Execution des diagnostics du Doctor..."
+  doctor_healthy: "Doctor: Tous les checks ont reussi"
+  doctor_degraded: "Doctor: Problemes trouves (voir rapport)"
+  doctor_fix_applied: "Doctor: Applique(s) {count} auto-fix(es)"
+
 errors:
   session_corrupted: "Le fichier de session semble corrompu. Tentative de recuperation..."
   handoff_missing: "Passation non trouvee. Utilisation de session.yaml + CLAUDE.md comme repli."

package/framework/i18n/pt.yaml

@@ -29,8 +29,8 @@ installer:
   created_overlays: "Criados diretorios de overlay de provedores"
   will_install: "Sera instalado:"
   agents_count: "13 definicoes de agentes (fases DISCOVER, PLAN, BUILD, DEPLOY)"
-  workflows_count: "6 blueprints de workflow"
-  templates_count: "6 templates (PRD, PRD Brownfield, Arquitetura, Task, QA Gate, Quick Brief)"
+  workflows_count: "7 blueprints de workflow"
+  templates_count: "10 templates (PRD, PRD Brownfield, Arquitetura, Task, QA Gate, Quick Brief, Brandbook, Design Tokens, Component Spec, Icon System)"
   constitution: "Constituicao (19 Artigos + Preambulo)"
   session_mgmt: "Sistema de gerenciamento de sessao"
   quality_gates: "Quality gates (veredictos em 3 niveis)"
@@ -85,6 +85,15 @@ status:
   backlog_items: "Itens do Backlog"
   high_priority: "alta prioridade"
 
+features:
+  brandbook_step: "Fase 0: Estabelecendo identidade de marca (Brandbook)"
+  tech_preset_loaded: "Tech preset carregado: {preset}"
+  tech_preset_detected: "Stack tecnologico detectado: {stack}"
+  doctor_running: "Executando diagnosticos do Doctor..."
+  doctor_healthy: "Doctor: Todos os checks passaram"
+  doctor_degraded: "Doctor: Problemas encontrados (veja relatorio)"
+  doctor_fix_applied: "Doctor: Aplicado(s) {count} auto-fix(es)"
+
 errors:
   session_corrupted: "Arquivo de sessao parece corrompido. Tentando recuperacao..."
   handoff_missing: "Handoff nao encontrado. Usando session.yaml + CLAUDE.md como fallback."

package/framework/intelligence/context-engine.md

@@ -2,7 +2,7 @@
 
 ## Purpose
 
-The Context Engine monitors context window usage and adapts what information is injected into each agent's prompt. As the context fills, fewer layers are injected to preserve quality for critical operations.
+The Context Engine monitors context window usage and adapts what information is injected into each agent's prompt. As the context fills, fewer layers are injected to maintain focus, while per-layer reinforcement INCREASES to compensate for the model forgetting initial instructions (Progressive Reinforcement Model).
 
 ---
 
@@ -12,16 +12,16 @@ Four brackets define behavior based on remaining context:
 
 | Bracket | Context Remaining | Behavior |
 |---------|-------------------|----------|
-| **FRESH** | 60-100% | All protocols active. Minimal injection — context is plentiful. |
-| **MODERATE** | 40-60% | Normal protocols. Summarize long outputs. Standard memory retrieval. |
-| **DEPLETED** | 25-40% | Reinforce Constitution + Agent scope. Skip optional patterns. Recover context via memory. |
-| **CRITICAL** | <25% | Constitution + Agent scope only. Trigger handoff to new session. Full memory dump for continuity. |
+| **FRESH** | 60-100% | All 6 layers active. Minimal reinforcement (1.5%) — context is plentiful. No memory injection. |
+| **MODERATE** | 40-60% | 5 layers (skip L4). Growing reinforcement (2.5%) — model starting to forget. Metadata memory. |
+| **DEPLETED** | 25-40% | 3 layers (L0-L2). Heavy reinforcement (4.0%) — significant context loss. Chunk memory recovery. |
+| **CRITICAL** | <25% | 2 layers (L0-L1). Maximum reinforcement (5.0%) — last interactions before handoff. Full memory dump. |
 
 ---
 
 ## Layered Context Injection
 
-The orchestrator injects context through 5 hierarchical layers:
+The orchestrator injects context through 6 hierarchical layers:
 
 | Layer | Name | Source | When Active |
 |-------|------|--------|-------------|
@@ -30,15 +30,18 @@ The orchestrator injects context through 5 hierarchical layers:
 | **L2** | Agent Scope | `chati.dev/agents/{agent}/` — mission, inputs, outputs, criteria | When agent is active |
 | **L3** | Pipeline State | `.chati/session.yaml` — pipeline position, scores, backlog | When session is active |
 | **L4** | Task Context | Active artifact + previous agent's handoff | When task is active |
+| **L5** | Keywords | `chati.dev/domains/keywords/` — dynamic rules from user prompt | FRESH + MODERATE only |
 
 ### Layer Activation by Bracket
 
-| Bracket | Active Layers | Approx. Token Budget |
-|---------|--------------|---------------------|
-| FRESH | L0, L1, L2, L3, L4 | ~2500 tokens |
-| MODERATE | L0, L1, L2, L3, L4 | ~2000 tokens |
-| DEPLETED | L0, L1, L2 | ~1500 tokens |
-| CRITICAL | L0, L1 | ~800 tokens |
+| Bracket | Active Layers | Budget Ratio | Claude (200K) | Gemini (1M) | Codex (128K) |
+|---------|--------------|-------------|---------------|-------------|--------------|
+| FRESH | L0, L1, L2, L3, L4, L5 | 1.5% | 3,000 | 15,000 | 1,920 |
+| MODERATE | L0, L1, L2, L3, L5 | 2.5% | 5,000 | 25,000 | 3,200 |
+| DEPLETED | L0, L1, L2 | 4.0% | 8,000 | 40,000 | 5,120 |
+| CRITICAL | L0, L1 | 5.0% | 10,000 | 50,000 | 6,400 |
+
+Budgets are proportional to the provider's context window. As context degrades, fewer layers are active (focus) but each active layer receives MORE reinforcement tokens (strength).
 
 ---
 
@@ -145,12 +148,14 @@ If Smart Continuation is insufficient, the orchestrator spawns a new session:
 
 ## Integration with Memory Layer
 
+Memory injection follows the same progressive reinforcement principle: as context degrades, memory becomes MORE important.
+
 | Bracket | Memory Level | Behavior |
 |---------|-------------|----------|
-| FRESH | None | Context is sufficient — no memory injection |
-| MODERATE | L1 Metadata | Light reminder of relevant memories (~50 tokens) |
-| DEPLETED | L2 Chunks | Context recovery via memory summaries (~200 tokens) |
-| CRITICAL | L3 Full | Full memory dump for session handoff (~1000+ tokens) |
+| FRESH | none | Context is sufficient — no memory injection needed |
+| MODERATE | metadata | Light reminder of relevant memories (~50 tokens) |
+| DEPLETED | chunks | Context recovery via memory summaries (~200 tokens) |
+| CRITICAL | full | Full memory dump for session handoff (~1000+ tokens) |
 
 ---
 
@@ -160,7 +165,7 @@ If Smart Continuation is insufficient, the orchestrator spawns a new session:
 
 ---
 
-*Context Engine v1.0 — Chati.dev Intelligence Layer*
+*Context Engine v2.0 — Chati.dev Intelligence Layer (Progressive Reinforcement Model)*
 
 ---
 

package/framework/presets/nextjs.yaml

@@ -0,0 +1,41 @@
+# Tech Preset: Next.js
+# Auto-loaded by Architect agent when Next.js stack is detected
+
+stack: nextjs
+displayName: "Next.js"
+
+detection:
+  - next.config.js
+  - next.config.mjs
+  - next.config.ts
+
+conventions:
+  routing: "App Router preferred (app/ directory)"
+  components: "Server Components by default, 'use client' for interactivity"
+  data_fetching: "Server-side fetch in page/layout, React Query for client"
+  styling: "Tailwind CSS or CSS Modules (no global CSS imports in components)"
+  api: "Route Handlers in app/api/ (not pages/api/)"
+
+structure:
+  - "app/ — Routes, layouts, and pages"
+  - "app/api/ — Route Handlers (REST endpoints)"
+  - "components/ — Reusable UI components"
+  - "lib/ — Utility functions and shared logic"
+  - "public/ — Static assets"
+
+rules:
+  - id: nextjs-ssr
+    text: "Prefer Server Components. Only add 'use client' when interactivity is needed."
+    priority: high
+  - id: nextjs-api
+    text: "Use Route Handlers (app/api/) for API endpoints. Not pages/api/."
+    priority: high
+  - id: nextjs-layout
+    text: "Use layout.tsx for shared UI. Avoid prop drilling through layouts."
+    priority: normal
+  - id: nextjs-metadata
+    text: "Export metadata from page.tsx for SEO. Use generateMetadata for dynamic pages."
+    priority: normal
+  - id: nextjs-loading
+    text: "Use loading.tsx and error.tsx for streaming and error boundaries."
+    priority: normal

package/framework/presets/node-express.yaml

@@ -0,0 +1,39 @@
+# Tech Preset: Node.js + Express
+# Auto-loaded by Architect agent when Express stack is detected
+
+stack: node-express
+displayName: "Node.js + Express"
+
+detection:
+  - "package.json containing express"
+
+conventions:
+  routing: "Express Router with modular route files"
+  middleware: "Error-handling middleware at the end of the chain"
+  validation: "Zod or Joi for request validation"
+  auth: "JWT or session-based, middleware pattern"
+  errors: "Centralized error handler with error classes"
+
+structure:
+  - "src/ — Application source code"
+  - "src/routes/ — Express route definitions"
+  - "src/controllers/ — Request handlers"
+  - "src/middleware/ — Custom middleware"
+  - "src/services/ — Business logic"
+  - "src/models/ — Data models (ORM or raw queries)"
+  - "src/utils/ — Utility functions"
+  - "src/config/ — Configuration and environment"
+
+rules:
+  - id: express-error-handler
+    text: "Always use a centralized error-handling middleware. Never swallow errors silently."
+    priority: critical
+  - id: express-validation
+    text: "Validate all request input (body, params, query) before processing."
+    priority: high
+  - id: express-async
+    text: "Wrap async route handlers to catch rejected promises (express-async-errors or manual try/catch)."
+    priority: high
+  - id: express-env
+    text: "Use environment variables for all configuration. Never hardcode secrets."
+    priority: critical

package/framework/presets/react-vite.yaml

@@ -0,0 +1,37 @@
+# Tech Preset: React + Vite
+# Auto-loaded by Architect agent when React+Vite stack is detected
+
+stack: react-vite
+displayName: "React + Vite"
+
+detection:
+  - vite.config.js
+  - vite.config.ts
+  - vite.config.mjs
+
+conventions:
+  routing: "React Router v6+ with createBrowserRouter"
+  components: "Functional components with hooks"
+  state: "React Context for simple state, Zustand/Jotai for complex"
+  styling: "Tailwind CSS or CSS Modules"
+  build: "Vite with ESBuild for dev, Rollup for production"
+
+structure:
+  - "src/ — Application source code"
+  - "src/components/ — Reusable UI components"
+  - "src/pages/ — Route-level components"
+  - "src/hooks/ — Custom React hooks"
+  - "src/lib/ — Utility functions"
+  - "src/assets/ — Static assets (images, fonts)"
+  - "public/ — Public static files"
+
+rules:
+  - id: vite-env
+    text: "Use import.meta.env for environment variables. Prefix with VITE_ for client exposure."
+    priority: high
+  - id: vite-lazy
+    text: "Use React.lazy() and Suspense for route-level code splitting."
+    priority: normal
+  - id: vite-alias
+    text: "Configure path aliases in vite.config.ts (e.g., @ for src/)."
+    priority: normal