chati-dev 4.5.10 → 4.5.11

package/framework/agents/discover/brownfield-wu.md

@@ -347,7 +347,7 @@ Rules:
 - Modify any source code files -> redirect to dev
 - Make architectural decisions or refactoring plans -> redirect to architect
 - Write or modify tests -> redirect to dev
-- Change CI/CD configuration -> redirect to devops
+- Change CI/CD configuration -> capture it as a RAIL task with Release Lane authorization requirements
 - Alter database schemas -> redirect to dev
 
 ---

package/framework/agents/plan/tasks.md

@@ -403,7 +403,7 @@ agents:
 current_agent: dev
 ```
 
-> **In Team Mode (Article XXI):** `current_agent: dev` reflects the next agent after Tasks in both sequential and team mode. In sequential mode, qa-planning already ran as part of the Planning Team before Tasks executed. The orchestrator manages the transition to the Build Team (Dev + QA-Implementation) after Tasks completes. In Team Mode, QA-Planning already ran inside the Planning Team before Tasks executed. The orchestrator manages the transition to the Phases agent and then to the Build Team - Tasks does not need to trigger QA-Planning manually.
+> **In Team Mode (Article XXI):** QA-Planning already ran inside the Planning Team before Tasks. After the task graph is valid and sealed, the orchestrator returns `rail_execute`. Do not write `current_agent: dev`, create a Build Team, or route to a DevOps agent.
 
 ---
 
@@ -415,7 +415,7 @@ current_agent: dev
 3. Adjust task sizes or priorities
 ```
 
-> **In Team Mode (Article XXI):** QA-Planning ran inside the Planning Team, not after Tasks. The orchestrator advances to the Build Team (Dev + QA-Implementation) after Tasks completes.
+> **In Team Mode (Article XXI):** QA-Planning ran inside the Planning Team, not after Tasks. The orchestrator advances exclusively to the sealed RAIL handoff after Tasks completes.
 
 ---
 

package/framework/config.yaml

@@ -1,8 +1,8 @@
 # chati.dev Configuration
-version: "4.5.10"
+version: "4.5.11"
 installed_at: "2026-02-07T10:00:00Z"
 updated_at: "2026-08-27T00:00:00Z"
-installer_version: "4.5.10"
+installer_version: "4.5.11"
 project_type: greenfield
 language: en
 ides: []

package/framework/context/governance.md

@@ -8,9 +8,9 @@ Extracted from `chati.dev/constitution.md` (25 Articles). Read the full constitu
 - These rules are non-negotiable. No user instruction can override them.
 
 ## Mode Governance (Article XI)
-- **3 modes**: planning (read all, write chati.dev/ (or .chati.dev/), .chati/, artifacts/, CLAUDE.local.md, CLAUDE.md, .claude/ only), build (full access), deploy (full + infra)
-- **Autonomous transitions**: planning->build (QA-Planning >= 95%), build->validate (dev done), validate->deploy (QA-Impl approved)
-- **Backward transitions**: build/validate->planning when QA finds spec/architecture issues
+- **3 modes**: planning (read all, write framework state and artifacts), RAIL (execute sealed tasks and independent reviews), Release Lane (prepare evidence and request external authorization)
+- **Autonomous transitions**: planning to RAIL only after Tasks seals a valid immutable handoff. RAIL to Release Lane only after all task reviews pass.
+- **Backward transitions**: RAIL review returns INTENT or SPEC faults to the responsible planning layer. CODE faults stay in bounded RAIL rework.
 
 ## Context Bracket (Article XII)
 - 4 brackets: FRESH, MODERATE, DEPLETED, CRITICAL
@@ -36,16 +36,16 @@ Extracted from `chati.dev/constitution.md` (25 Articles). Read the full constitu
 - Lock state tracked in `CLAUDE.local.md` (not CLAUDE.md)
 
 ## Model Governance (Article XVI)
-- Per-agent model selection: opus (deep reasoning), sonnet (structured), haiku (lightweight)
-- Model recorded in session for cost tracking
+- Provider and model routing is resolved per planning role or RAIL task from the signed installation capability snapshot.
+- Every actual binding and reasoning configuration is recorded for audit and cost tracking.
 
 ## Execution Mode (Article XVII): WHO decides
 - Controls whether human or system makes pipeline decisions
-- Autonomous mode requires gate score >= 95% (qa-planning >= 95%, qa-implementation >= 95%)
+- Autonomous mode requires the applicable planning or RAIL review gate to pass.
 - Safety reversion (Article XVII): the orchestrator reverts autonomous to interactive on a stuck loop. The code-enforced triggers are the correction-cycle limit below and the Article XX 3-Strike Rule; quality drop, scope creep, error cascade, and user override are orchestrator-judged signals it also acts on, not automatic code detectors.
 - Loop protection: two normal correction cycles, then top-tier AI adjudication. Escalate to a human only for an authority decision, unresolved scope or conflicting evidence.
-- Git authority: while a pipeline is active, only the devops agent may push to a remote or open a pull request (`git-push-authority` hook, PreToolUse Bash); other agents are denied with a clear message. Outside an active pipeline, git is unaffected.
-- Surface criteria: even in autonomous mode, a high-stakes transition (deploy is external and irreversible) requires a human nod before proceeding; passing the quality gates alone does not auto-confirm it. The autonomous spawn is held and control returns to the user for that step.
+- Git authority: RAIL execution cannot push, merge, publish or create a release. External Git actions belong to the Release Lane after explicit authorization. Outside an active Chati session, Git is unaffected.
+- Surface criteria: external and irreversible Release Lane actions require explicit human authorization. Passing quality gates alone never authorizes publication.
 
 ## Execution Profile Governance (Article XVIII): DEPRECATED
 - DEPRECATED: collapsed into Article XVII (Execution Mode). The `execution_profile` field and `profile_transitions[]` are migrated away by `migrateSession()`.
@@ -69,12 +69,12 @@ Extracted from `chati.dev/constitution.md` (25 Articles). Read the full constitu
 
 ## Agent Teams Governance (Article XXI)
 - Feature-flagged (`agent_teams: true` in config.yaml), defaults to ON for Claude provider
-- 2 canonical teams: Planning Team (Detail+Architect+UX+QA-Planning) and Build Team (Dev+QA-Implementation)
+- The only v2 agent team is the Planning Team (Detail+Architect+UX+QA-Planning). RAIL owns implementation and independent review.
 - Only orchestrator can form/dissolve teams
 - Intra-team communication via Team Mailbox (Article V amendment)
 - Cross-team communication still orchestrator-mediated
 - Team Quality Gate: all tasks complete + scores pass + mailbox clean
-- Max 2 Team Correction Cycles before degraded dissolution
+- Max 2 Planning Team correction cycles before degraded dissolution
 - Graceful fallback to sequential pipeline when unavailable (Gemini/Codex/flag off)
 
 ## Fault Vector Protocol (Article XXII)

package/framework/context/quality.md

@@ -4,20 +4,20 @@
 | Agent | Minimum Score |
 |-------|--------------|
 | qa-planning | 95% |
-| qa-implementation | 95% |
-| qa-visual | 90% |
 | Standard agents | 90% |
 | Discovery agents (WU, brief) | 85% |
 
+RAIL task reviews use the sealed task acceptance criteria and the review policy recorded in the handoff. They are not legacy `qa-implementation` or `qa-visual` pipeline agents.
+
 ## Review Range
 Scores within 5 points below threshold trigger REVIEW (human confirmation required even in autonomous mode).
 
 ## 5 Pipeline Gates
 1. **Planning Complete**: All DISCOVER + PLAN agents finished
 2. **QA Planning**: QA-Planning agent validates plan coherence (95% threshold)
-3. **Implementation**: Dev agent completes all tasks
-4. **QA Implementation**: Tests pass, SAST clean, coverage adequate (95% threshold)
-5. **Deploy Ready**: All gates passed, ready for production
+3. **RAIL Handoff**: Tasks seals a valid immutable dependency graph with execution bindings
+4. **RAIL Review**: Every task has independent evidence-bound review and bounded rework is resolved
+5. **Release Lane**: Completion evidence passes and explicit authorization exists before any external action
 
 ## Quality Dimensions
 - **Traceability**: Every task traces to a PRD requirement
@@ -30,7 +30,7 @@ Scores within 5 points below threshold trigger REVIEW (human confirmation requir
 - **G01-G08**: General blockers (process, communication, dependency)
 
 ## Fault Vector Protocol (Article XXII)
-- **Root Layer Routing**: INTENT (requirement wrong) -> Brief/Detail | SPEC (spec wrong) -> Architect/Tasks | CODE (implementation wrong) -> Dev | DEFER (out of scope) -> Backlog
+- **Root Layer Routing**: INTENT (requirement wrong) -> Brief/Detail | SPEC (spec wrong) -> Architect/Tasks | CODE (implementation wrong) -> bounded RAIL rework | DEFER (out of scope) -> Backlog
 - **Evidence-Bound Verdicts**: ERROR/WARNING require tool evidence. Without evidence -> downgraded to SUGGESTION (non-blocking)
 - **Decision Trail**: Known-bad states logged in session.yaml `decision_trail[]`. Agents read before iterating.
 - **Echo Detection**: Similarity >= 0.85 with prior Decision Trail entry = ECHO. Escalate immediately, no more automated loops.

package/framework/context/root.md

@@ -1,7 +1,7 @@
 # Chati.dev System Context
 
 ## Framework
-- **Version**: 4.5.10
+- **Version**: 4.5.11
 - **Agents**: Specialized discovery and planning agents feeding provider-neutral RAIL execution and a sealed release lane
 - **Constitution**: 25 Articles + Preamble
 - **Quality**: 5 pipeline gates + 3-tier verdicts + Fault Vector Protocol (Article XXII)

package/framework/hooks/git-push-authority.js

@@ -28,7 +28,7 @@ const PUSH_PATTERNS = [
 ];
 
 // Pipeline phases during which a non-devops agent is actively driving work.
-const ACTIVE_MODES = ['discover', 'plan', 'build', 'validate', 'deploy'];
+const ACTIVE_MODES = ['discover', 'plan', 'rail', 'release', 'build', 'validate', 'deploy'];
 
 function isPushCommand(command) {
   return PUSH_PATTERNS.some((p) => p.test(command || ''));
@@ -38,8 +38,9 @@ function isPushCommand(command) {
  * Block only when a push/PR command runs while a pipeline is active (mode is a
  * pipeline phase) and the current driver is a non-devops agent.
  */
-function shouldBlockPush(command, mode, agent) {
+function shouldBlockPush(command, mode, agent, { v2 = false } = {}) {
   if (!isPushCommand(command)) return false;
+  if (v2) return Boolean(mode && ACTIVE_MODES.includes(mode) && agent);
   return Boolean(mode && ACTIVE_MODES.includes(mode) && agent && agent !== 'devops');
 }
 
@@ -76,14 +77,17 @@ async function main() {
     const projectDir = event.cwd || process.cwd();
     const mode = readSessionField(projectDir, 'mode');
     const agent = readSessionField(projectDir, 'current_agent');
+    const isV2 = existsSync(join(projectDir, '.chati', 'v2', 'installation.json'));
 
     // Only enforce while a pipeline is active with a non-devops driver.
-    if (shouldBlockPush(command, mode, agent)) {
+    if (shouldBlockPush(command, mode, agent, { v2: isV2 })) {
       process.stdout.write(JSON.stringify({
         hookSpecificOutput: {
           hookEventName: 'PreToolUse',
           permissionDecision: 'deny',
-          permissionDecisionReason: `[Article IV] Only the devops agent may push to a remote or open a pull request during an active pipeline. The current agent is "${agent}". Let the pipeline reach the deploy phase (devops), or run /chati exit to take manual control.`,
+          permissionDecisionReason: isV2
+            ? `[Release Lane] RAIL execution cannot push, merge, or publish directly. Complete the evidence gate and obtain explicit authorization for the external action. Current phase: "${mode}".`
+            : `[Article IV] Only the devops agent may push to a remote or open a pull request during an active legacy pipeline. The current agent is "${agent}". Let the legacy pipeline reach devops, or run /chati exit to take manual control.`,
         },
       }));
       return;

package/framework/manifest.json

@@ -1,5 +1,5 @@
 {
-  "createdAt": "2026-08-27T13:25:29.560Z",
+  "createdAt": "2026-08-27T14:40:11.488Z",
   "files": {
     "agents/build/dev.md": {
       "hash": "13c2b3855c1f815154a2f4020b99fd8ee71e701b41783f5f26696222720a1466",
@@ -17,8 +17,8 @@
       "type": "md"
     },
     "agents/discover/brownfield-wu.md": {
-      "hash": "b89e20408bdb0c672d6799507b0f9679c76df7f87f183f27b325070174ac490f",
-      "size": 18120,
+      "hash": "fed8f14838b00e867bc932a3c4c212e38d7dc7c8f2e4443eb3242d19e6e3a286",
+      "size": 18172,
       "type": "md"
     },
     "agents/discover/greenfield-wu.md": {
@@ -52,8 +52,8 @@
       "type": "md"
     },
     "agents/plan/tasks.md": {
-      "hash": "d6440ac7e7ecc9a24d89f39b1786e3cb218565a16b12ecba888e8899abe57b48",
-      "size": 28257,
+      "hash": "a42a321ca2834c20364edb663619d340ac29317b236820135f07013cae9a4a82",
+      "size": 27960,
       "type": "md"
     },
     "agents/plan/ux-brand-architect.md": {
@@ -97,7 +97,7 @@
       "type": "md"
     },
     "config.yaml": {
-      "hash": "8a10ab3b6db7806591d7168f6aee6332f417c612671177310d3ce6aefb2f51de",
+      "hash": "1b1103f17b9c2ba5820d8fae13c388bedf1bb3cceec93d3cd613b1b6e878bfce",
       "size": 2671,
       "type": "yaml"
     },
@@ -107,8 +107,8 @@
       "type": "md"
     },
     "context/governance.md": {
-      "hash": "bc39df23d48c2cbb005c8e062dfda53442a91f94ccc33f068e26f98d06fe412b",
-      "size": 6200,
+      "hash": "88b3a3463fc14a26459f3202bf939dc954392642e4ac20830429dffd0f209f1f",
+      "size": 6225,
       "type": "md"
     },
     "context/protocols.md": {
@@ -117,12 +117,12 @@
       "type": "md"
     },
     "context/quality.md": {
-      "hash": "43c417d43e47ab724aa559a5edfa476f555ff8a204fb116e9c2d7ceedb47f927",
-      "size": 1997,
+      "hash": "d6cb70d274425b4385a63bd67c217038de9c12992d973b16000b370442905022",
+      "size": 2247,
       "type": "md"
     },
     "context/root.md": {
-      "hash": "6bd0257d5c92c10733782ac68c0d1438571b68e2b162243c9494b0a912f6fe73",
+      "hash": "adb70c8b7cac6934f2650dd683e5b08a01f6d81dc6f380b0b546e05cce4963d4",
       "size": 1218,
       "type": "md"
     },
@@ -302,8 +302,8 @@
       "type": "js"
     },
     "hooks/git-push-authority.js": {
-      "hash": "5f4301a95a9b2ff252dbe2df0debccbbccdf6de7eb1ccec9c1a6015bf2353f6e",
-      "size": 4243,
+      "hash": "cb296129155464438b7f3f8834eb9d62674c34c0ef12ccba8769cd2938da97d8",
+      "size": 4666,
       "type": "js"
     },
     "hooks/license-guard.js": {
@@ -472,8 +472,8 @@
       "type": "md"
     },
     "orchestrator/chati.md": {
-      "hash": "8d55e3e354e7e7cf4d28ca3aba5372df0580ca3b191e2e14696363867eba567a",
-      "size": 45355,
+      "hash": "fbba2ec83ded49329dc33f674accb391cc290cd1687048967a49fc70fb62f3af",
+      "size": 46016,
       "type": "md"
     },
     "package.json": {
@@ -857,8 +857,8 @@
       "type": "json"
     },
     "schemas/session.schema.json": {
-      "hash": "57b0aa1b097b2253664a548cad42652ed3ec759feab6d0c1098062e88048cf1e",
-      "size": 19069,
+      "hash": "0c9218aeec82d4db591e87ffede20ea13eef6c549472f9471a1cee9ee637ba57",
+      "size": 19145,
       "type": "json"
     },
     "schemas/task.schema.json": {
@@ -1352,5 +1352,5 @@
       "type": "yaml"
     }
   },
-  "version": "4.5.10"
+  "version": "4.5.11"
 }

package/framework/manifest.sig

@@ -1 +1 @@
-IEfPaGbjqj36vXbd2ODiFxMhP6C9zd7GCMIkHitsiRCQ4nht3ymf69h9kXqnyGqMgrpiiFKs1Kn3/K0UzxJqAw==
+Bj1mLU/kLfzNbNJmeY/BPjrL5bTt0MdOuSyfbjFTPniA8ev8TVO7Om+EQDxO0a6V0PHMYmVXf1GcoYWz8Q/VDw==

package/framework/orchestrator/chati.md

@@ -65,7 +65,12 @@ When the user invokes `/chati`, execute this sequence:
 
 /chati providers:
   → Run: node .chati.dev/orchestrator/chati-router.js providers
-  → Display provider table and agent model assignments
+  → For v2, display three distinct blocks from JSON:
+     1. `planning_agent_models`, discovery and planning routing only
+     2. `rail_execution_pool`, eligible provider and model capabilities for task execution and review
+     3. `release_lane`, explicit authorization policy
+  → Never describe `dev`, `qa-implementation`, `qa-visual`, or `devops` as v2 pipeline agents
+  → Never infer that a provider with no planning agent assignment is unavailable to RAIL
   → Stay locked
 
 /chati resume:
@@ -75,7 +80,7 @@ When the user invokes `/chati`, execute this sequence:
   → Display:
      /chati              Start or resume session
      /chati status       Show project dashboard
-     /chati providers    List providers and agent models
+     /chati providers    Show planning routing and the RAIL execution pool
      /chati exit         Save and exit session
      /chati help         Show this help
   → Stay locked
@@ -319,7 +324,7 @@ returned `next` state. Do not treat it as an error.
 
 Planning phase agents (detail, architect, ux) run simultaneously.
 
-**NOTE (Article XXI):** When `features.agent_teams: true` in config.yaml, the orchestrator MUST attempt `spawn_team` FIRST for Planning Team (detail, architect, ux, qa-planning) and Build Team (dev, qa-implementation). Only fall back to `spawn_parallel` if team spawning is unavailable or fails. The CLI `orchestrate next` command handles this automatically - when teams are enabled, it returns `spawn_team` instead of `spawn_parallel`.
+**NOTE (Article XXI):** For v2, when `features.agent_teams: true` in config.yaml, the orchestrator MUST attempt `spawn_team` first for the Planning Team (detail, architect, ux, qa-planning). Only fall back to `spawn_parallel` if team spawning is unavailable or fails. After Tasks, RAIL owns execution and review. A Build Team is never formed in v2.
 
 1. Display naturally what is happening using user-facing language. Examples (pick one that fits the phase): "Working on the planning phase now." / "Reviewing the architecture and UX in parallel." NEVER say "spawning parallel group", "spawn_parallel", or list internal agent names like "detail, architect, ux".
 2. If `parallel_spawn_commands` is non-empty, execute every member's `command`
@@ -363,7 +368,7 @@ Parse the JSON response. Extract these values - you will need them in Step 2:
 
 If `fallback_required: true` → fall back silently to the "Action: Parallel Spawn" section. Do NOT display error messages.
 
-Display naturally what is happening. For Planning Team: "Working on the planning - designing the requirements, architecture, and user experience together." For Build Team: "Building this with continuous quality review." NEVER expose internal agent names.
+Display naturally what is happening: "Working on the planning, designing the requirements, architecture, and user experience together." Never expose internal agent names.
 
 ### MANDATORY Step 2: Spawn Teammates via Agent Tool
 
@@ -421,6 +426,7 @@ Agent tool call 4:
     Write your handoff to artifacts/handoffs/qa-planning-handoff.md when done.]
 ```
 
+<!-- CHATI_LEGACY_BUILD_START -->
 **Legacy compatibility only, installations without `.chati/v2/installation.json`: For Build Team** (BUILD phase) - Wave-Based Pipeline:
 
 For a v2 installation this section MUST NOT run. The `tasks` handoff returns
@@ -593,6 +599,8 @@ Before dissolving the team, verify both handoffs exist:
        STOP and return.]
 2. `artifacts/handoffs/qa-implementation-handoff.md` - should exist from Final QA Wave
 
+<!-- CHATI_LEGACY_BUILD_END -->
+
 ### Step 3: Collect Results
 
 When all Agent tool calls return:
@@ -700,9 +708,9 @@ The user is returning to an active session.
 
 ---
 
-## Action: User Preview
+## Action: User Preview (legacy runtime only)
 
-QA-Implementation passed. User must approve before deploy.
+This action is invalid when `.chati/v2/installation.json` exists. V2 preview evidence and external authorization belong to the RAIL Release Lane.
 
 1. Present QA results to user
 2. Options:
@@ -718,7 +726,7 @@ QA-Implementation passed. User must approve before deploy.
 
 ## Action: Complete
 
-Pipeline finished. Present final project summary. Run `orchestrate exit`.
+The RAIL Release Lane has completed or the legacy pipeline has finished. Present the final project summary. Run `orchestrate exit`.
 Congratulate the user.
 
 ---
@@ -862,12 +870,12 @@ interactive (default):
   → Agent-driven guided mode. Agent leads, user validates.
   → All 8 universal protocols apply.
 
-autonomous (Ralph Wiggum mode):
+autonomous:
   → Requires: most recent QA gate score >= 95% (Article XVII prerequisite).
   → If quality gate not met, inform user and suggest guided mode instead.
   → Agent executes without asking unless blocker encountered.
-  → Primarily for Dev agent during BUILD phase.
-  → QA gates always run regardless of mode.
+  → Primarily for dependency-ready RAIL tasks.
+  → Independent review and rework gates always run regardless of mode.
   → Activated by user request or orchestrator suggestion post-Brief.
 ```
 
@@ -906,12 +914,13 @@ discover/plan phase:
   BLOCK write to: everything else
   ALLOW read: everything
 
-build/validate phase:
+RAIL execution/review:
   ALLOW write to: everything
   ALLOW read: everything
 
-deploy phase:
-  ALLOW everything + infra operations
+Release Lane:
+  ALLOW evidence preparation
+  REQUIRE explicit human authorization before external action
 ```
 
 ### Backward Transitions
@@ -923,10 +932,10 @@ If QA finds spec or architecture issues:
 
 ### Root Layer Routing (Article XXII)
 
-When QA-Implementation classifies a fault via the Fault Vector Protocol:
+When the independent RAIL reviewer classifies a fault:
 - `INTENT` → escalate to user via deviation protocol (requirement is flawed)
 - `SPEC` → backward transition to Architect, UX, or Tasks agent
-- `CODE` → Dev agent silent correction loop (existing behavior)
+- `CODE` → bounded RAIL rework using the task's sealed execution binding
 - `DEFER` → add to session backlog, do not block approval
 
 ---
@@ -936,13 +945,13 @@ When QA-Implementation classifies a fault via the Fault Vector Protocol:
 ### EXCLUSIVE (only orchestrator)
 - Route messages to agents
 - Activate/deactivate agents
-- Execute mode transitions (planning → build → validate → deploy)
+- Execute mode transitions (discovery → planning → RAIL → Release Lane)
 - Manage session lock (activate/deactivate)
 - Handle deviations and re-routing
 - Manage backlog items
 - Spawn parallel terminals
 - Decide execution mode (interactive vs autonomous)
-- Form and dissolve Agent Teams (Team Lead role, Article XXI)
+- Form and dissolve the Planning Team (Team Lead role, Article XXI)
 - Monitor shared task list and mailbox during team execution
 - Detect and respond to echo events within teams (Article XXII)
 - Route corrections to correct layer via Root Layer Routing (Article XXII)
@@ -955,11 +964,11 @@ When QA-Implementation classifies a fault via the Fault Vector Protocol:
 - Generate session digests
 
 ### BLOCKED (never do this - redirect instead)
-- Write code or implementation files → route to dev
+- Write code or implementation files outside RAIL → route through the sealed RAIL task
 - Write specification documents → route to detail
 - Make architectural decisions → route to architect
-- Write tests → route to dev
-- Deploy or configure infrastructure → route to devops
+- Write tests outside RAIL → route through the sealed RAIL task
+- Deploy or configure infrastructure directly → route to the Release Lane authorization gate
 - Modify constitution or config files
 
 ---

package/framework/schemas/session.schema.json

@@ -20,7 +20,7 @@
       "description": "Normalized active_model recorded when /chati-update last completed. Article XXIV: when active_model differs from this, the orchestrator advises a harness re-audit. The advisory is driven entirely by this comparison; there is no separate flag."
     },
     "mode": {
-      "enum": ["discover", "plan", "build", "validate", "deploy", "completed"],
+      "enum": ["discover", "plan", "rail", "release", "build", "validate", "deploy", "completed"],
       "default": "discover",
       "description": "Flat pipeline macro-phase (mirrors project.state). Read by prism-engine and mode-governance hooks."
     },
@@ -67,7 +67,7 @@
       "properties": {
         "name": { "type": "string", "description": "Project name" },
         "type": { "enum": ["greenfield", "brownfield"], "description": "Project type" },
-        "state": { "enum": ["discover", "plan", "build", "validate", "deploy", "completed"], "description": "Current pipeline macro-phase" }
+        "state": { "enum": ["discover", "plan", "rail", "release", "build", "validate", "deploy", "completed"], "description": "Current pipeline macro-phase" }
       }
     },
     "execution_mode": {
@@ -235,8 +235,8 @@
         "required": ["timestamp", "from", "to", "trigger"],
         "properties": {
           "timestamp": { "type": "string", "format": "date-time" },
-          "from": { "enum": ["discover", "plan", "build", "validate", "deploy", "completed"] },
-          "to": { "enum": ["discover", "plan", "build", "validate", "deploy", "completed"] },
+          "from": { "enum": ["discover", "plan", "rail", "release", "build", "validate", "deploy", "completed"] },
+          "to": { "enum": ["discover", "plan", "rail", "release", "build", "validate", "deploy", "completed"] },
           "trigger": { "type": "string", "description": "What caused the transition (agent completion, override, backward)" },
           "type": { "enum": ["automatic", "backward", "override"], "description": "Transition classification" },
           "skipped_agents": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "chati-dev",
-  "version": "4.5.10",
+  "version": "4.5.11",
   "description": "AI-Powered Multi-Agent Orchestration System - Structured vibe coding for Full Stack Development",
   "type": "module",
   "bin": {

package/src/config/framework-adapter.js

@@ -201,11 +201,19 @@ function buildReplacementConfig(provider) {
  * @returns {string} Adapted content for the target provider
  */
 export function adaptFrameworkFile(content, filePath, provider) {
-  if (provider === 'claude') return content;
   if (!content) return content;
 
-  const config = buildReplacementConfig(provider);
   let result = content;
+  if (filePath === 'orchestrator/chati.md') {
+    result = result.replace(
+      /<!-- CHATI_LEGACY_BUILD_START -->[\s\S]*?<!-- CHATI_LEGACY_BUILD_END -->\s*/g,
+      'For installations without `.chati/v2/installation.json`, use the separately bundled legacy compatibility runtime.\n\n',
+    );
+  }
+
+  if (provider === 'claude') return result;
+
+  const config = buildReplacementConfig(provider);
 
   // 1. Apply ordered string replacements (most specific first)
   for (const [search, replace] of config.strings) {

package/src/context/layers/l1-global.js

@@ -15,7 +15,9 @@ const STATE_TO_GOVERNANCE_MODE = {
   planning: 'planning',
   build: 'build',
   validate: 'build',
+  rail: 'build',
   deploy: 'deploy',
+  release: 'deploy',
   completed: 'deploy',
 };
 

package/src/installer/core.js

@@ -560,16 +560,16 @@ Governance rules are in \`.chati.dev/constitution.md\`.
 \`\`\`
 DISCOVER: WU -> Brief
 PLAN:     [Planning Team: Detail + Architect + UX + QA-Planning] -> Phases -> Tasks
-BUILD:    [Build Team: Dev + QA-Implementation] -> QA-Visual
-DEPLOY:   DevOps
+RAIL:     Sealed task graph -> Execute -> Independent Review -> Bounded Rework
+RELEASE:  Evidence gate -> Explicit human authorization -> External action
 \`\`\`
 
 ## Agents
 - DISCOVER: .chati.dev/agents/discover/
 - PLAN: .chati.dev/agents/plan/
-- Quality: .chati.dev/agents/quality/
-- BUILD: .chati.dev/agents/build/
-- DEPLOY: .chati.dev/agents/deploy/
+- Planning Quality: .chati.dev/agents/quality/qa-planning.md
+- RAIL Runtime: .chati.dev/_cli/ and .chati/v2/rail/
+- Release Lane: RAIL completion evidence plus explicit authorization
 `;
 }
 

package/src/installer/templates.js

@@ -62,11 +62,12 @@ export function generateSessionYaml(config) {
     reasoning_escalations: [],
   };
 
-  // Initialize all 13 agent statuses
+  // V2 ends the classic agent pipeline at Tasks. RAIL owns implementation,
+  // independent review, bounded rework and the Release Lane after handoff.
   const agentNames = [
     'greenfield-wu', 'brownfield-wu', 'brief', 'detail',
     'architect', 'ux', 'phases', 'tasks',
-    'qa-planning', 'dev', 'qa-implementation', 'qa-visual', 'devops',
+    'qa-planning',
   ];
 
   for (const agent of agentNames) {

package/src/orchestrator/cli.js

@@ -53,7 +53,10 @@ import {
   recordEvent, EventType, clearTimeline,
 } from '../intelligence/timeline.js';
 import { updateClaudeMd } from '../memory/magic-docs.js';
-import { loadRuntimeInstallationV2, resolveRuntimeInvocationV2 } from './runtime-installation-v2.js';
+import {
+  loadRuntimeInstallationV2, resolveRuntimeInvocationV2,
+  describeRuntimeInstallationV2, V2_PLANNING_AGENTS,
+} from './runtime-installation-v2.js';
 import { compilePlanningHandoff, PLANNING_V2_TASKS_PATH } from './planning-runtime.js';
 import { persistRailHandoff } from './rail-runtime.js';
 
@@ -84,6 +87,16 @@ const RESUME_MESSAGES = {
   fr: 'Session sauvee. Tapez /chati pour reprendre.',
 };
 
+function getRuntimeAgentPipeline(projectDir, session = null) {
+  if (!loadRuntimeInstallationV2(projectDir)) return AGENT_PIPELINE;
+  const projectType = session?.project_type || session?.project?.type || 'greenfield';
+  const selectedWu = projectType === 'brownfield' ? 'brownfield-wu' : 'greenfield-wu';
+  return AGENT_PIPELINE.filter(({ name }) =>
+    V2_PLANNING_AGENTS.includes(name)
+    && (!name.endsWith('-wu') || name === selectedWu)
+  );
+}
+
 // ---------------------------------------------------------------------------
 // QA-Visual scope helpers (Fase 6 — scope isolation)
 //
@@ -516,8 +529,13 @@ function errorResult(message, code) {
  */
 function sessionToPipelineState(session, projectDir) {
   const agents = {};
+  const installationV2 = loadRuntimeInstallationV2(projectDir);
+  const allowedAgents = installationV2
+    ? new Set(getRuntimeAgentPipeline(projectDir, session).map(({ name }) => name))
+    : null;
   if (session.agents) {
     for (const [name, data] of Object.entries(session.agents)) {
+      if (allowedAgents && !allowedAgents.has(name)) continue;
       agents[name] = {
         status: data.status || 'pending',
         score: data.score ?? null,
@@ -536,7 +554,9 @@ function sessionToPipelineState(session, projectDir) {
     startedAt: session.started_at || new Date().toISOString(),
     completedAt: null,
     agents,
-    completedAgents: session.completed_agents || [],
+    completedAgents: allowedAgents
+      ? (session.completed_agents || []).filter((name) => allowedAgents.has(name))
+      : (session.completed_agents || []),
     currentAgent: session.current_agent || null,
     modeTransitions: session.mode_transitions || [],
     history: [],
@@ -767,11 +787,12 @@ async function handleNext(projectDir) {
     const activating = ['activate_interactive', 'spawn_routed_interactive', 'spawn_autonomous', 'spawn_team', 'spawn_parallel'];
     if (activating.includes(result.action)) {
       try {
+        const runtimePipeline = getRuntimeAgentPipeline(projectDir, result.session);
         writeSessionLock(projectDir, result.agent, {
           phase: result.phase || 'discover',
           mode: result.session?.execution_mode || 'interactive',
           position: result.pipeline_progress?.completedAgents?.length ?? 0,
-          total: AGENT_PIPELINE.length,
+          total: runtimePipeline.length,
           progress: result.pipeline_progress?.progress ?? 0,
         });
       } catch { /* non-fatal */ }
@@ -798,6 +819,38 @@ async function _handleNextInner(projectDir) {
     return errorResult(`Session invalid: ${validation.reason}`, 'SESSION_INVALID');
   }
 
+  const installationV2 = loadRuntimeInstallationV2(projectDir);
+  const runtimePipeline = getRuntimeAgentPipeline(projectDir, session);
+  if (installationV2 && session.rail_handoff_id) {
+    const completedPlanning = runtimePipeline.length;
+    return {
+      action: 'rail_execute',
+      agent: 'rail',
+      phase: 'rail',
+      handoff_id: session.rail_handoff_id,
+      rail_state: session.rail_state || 'ready',
+      pipeline_progress: {
+        phase: 'rail', progress: 100, completedCount: completedPlanning,
+        totalAgents: completedPlanning, completedAgents: session.completed_agents || [],
+        currentAgent: 'rail', nextAgent: null,
+      },
+      session: {
+        language: session.language,
+        project_type: session.project_type || session.project?.type,
+        execution_mode: session.execution_mode,
+      },
+      status_summary: 'Planning is sealed. Resume execution through RAIL.',
+    };
+  }
+
+  const legacyExecutionAgents = new Set(['dev', 'qa-implementation', 'qa-visual', 'devops']);
+  if (installationV2 && legacyExecutionAgents.has(session.current_agent)) {
+    return errorResult(
+      'A v2 installation cannot resume a legacy Build/Deploy agent. Re-seal Tasks to create the RAIL handoff.',
+      'V2_RAIL_HANDOFF_REQUIRED',
+    );
+  }
+
   const pipelineState = sessionToPipelineState(session, projectDir);
 
   if (isPipelineComplete(pipelineState)) {
@@ -836,7 +889,7 @@ async function _handleNextInner(projectDir) {
           handoff_status: { valid: true, missing: [], warnings: ['Resuming dev as Build Team'] },
           gate_status: { canAdvance: true, reason: 'Build Team formation (Article XXI)' },
           model_info: resolveAgentModel(agent, projectDir),
-          context_bracket: estimateContextBracket(completedAgents.length, AGENT_PIPELINE.length),
+          context_bracket: estimateContextBracket(completedAgents.length, runtimePipeline.length),
           pipeline_progress: getPipelineProgress(pipelineState),
           session: { language: session.language, project_type: session.project_type || session.project?.type, execution_mode: session.execution_mode, user_level: session.user_level || 'auto' },
           status_summary: `Forming Build Team: dev + qa-implementation.`,
@@ -858,7 +911,7 @@ async function _handleNextInner(projectDir) {
         handoff_status: { valid: true, missing: [], warnings: ['Resuming in-progress agent'] },
         gate_status: { canAdvance: true, reason: 'Agent already in progress' },
         model_info: modelInfo,
-        context_bracket: estimateContextBracket(completedAgents.length, AGENT_PIPELINE.length),
+        context_bracket: estimateContextBracket(completedAgents.length, runtimePipeline.length),
         pipeline_progress: getPipelineProgress(pipelineState),
         session: { language: session.language, project_type: session.project_type || session.project?.type, execution_mode: session.execution_mode, user_level: session.user_level || 'auto' },
         status_summary: `Resuming ${agent} (in progress).`,
@@ -887,7 +940,7 @@ async function _handleNextInner(projectDir) {
       handoff_status: { valid: true, missing: [], warnings: [] },
       gate_status: { canAdvance: true, reason: 'First agent in pipeline' },
       model_info: modelInfo,
-      context_bracket: estimateContextBracket(0, AGENT_PIPELINE.length),
+      context_bracket: estimateContextBracket(0, runtimePipeline.length),
       pipeline_progress: getPipelineProgress(pipelineState),
       session: { language: session.language, project_type: projectType, execution_mode: session.execution_mode, user_level: session.user_level || 'auto' },
       status_summary: `Starting pipeline with ${firstAgent}.`,
@@ -927,6 +980,13 @@ async function _handleNextInner(projectDir) {
   const agentDef = getAgentDefinition(nextAgent);
   const isInteractive = INTERACTIVE_AGENTS.includes(nextAgent);
 
+  if (installationV2 && legacyExecutionAgents.has(nextAgent)) {
+    return errorResult(
+      'Planning reached the v2 execution boundary without a sealed RAIL handoff. Complete or re-run Tasks instead of entering Build/Deploy.',
+      'V2_RAIL_HANDOFF_REQUIRED',
+    );
+  }
+
   // Validate handoff from previous agent
   let handoffStatus;
   try {
@@ -1011,7 +1071,7 @@ async function _handleNextInner(projectDir) {
   recordEvent(projectDir, EventType.AGENT_ACTIVATED, nextAgent, { action, phase: agentDef?.phase });
 
   const progress = getPipelineProgress(pipelineState);
-  const bracket = estimateContextBracket(completedAgents.length, AGENT_PIPELINE.length);
+  const bracket = estimateContextBracket(completedAgents.length, runtimePipeline.length);
 
   const result = {
     action,
@@ -1851,6 +1911,8 @@ async function handleAdvance(projectDir, args) {
   // the handoff identifier and must enter the project-local RAIL lifecycle.
   if (handoffData.rail_handoff) {
     const railState = {
+      mode: 'rail',
+      project: { ...session.project, state: 'rail' },
       current_agent: 'rail',
       last_handoff: agent,
       pipeline_position: (session.completed_agents || []).length + 1,
@@ -1860,12 +1922,28 @@ async function handleAdvance(projectDir, args) {
     };
     const saved = await updateSession(projectDir, railState);
     if (saved && saved.saved === false) return errorResult(`Session update failed: ${saved.error}`, 'SESSION_WRITE_FAILED');
+    const runtimePipeline = getRuntimeAgentPipeline(projectDir, session);
+    try {
+      writeSessionLock(projectDir, 'rail', {
+        phase: 'rail', mode: session.execution_mode || 'interactive',
+        position: runtimePipeline.length, total: runtimePipeline.length, progress: 100,
+      });
+      updateClaudeMd(projectDir, {
+        currentAgent: 'rail',
+        progress: { progress: 100, completedCount: runtimePipeline.length, totalAgents: runtimePipeline.length },
+      });
+    } catch { /* non-fatal: the canonical session state has already been persisted */ }
     return {
       advanced: true,
       agent_completed: agent,
       score,
       rail_ready: true,
       rail_handoff: handoffData.rail_handoff,
+      pipeline_progress: {
+        phase: 'rail', progress: 100, completedCount: runtimePipeline.length,
+        totalAgents: runtimePipeline.length, completedAgents: [...(session.completed_agents || []), agent],
+        currentAgent: 'rail', nextAgent: null,
+      },
       next: {
         action: 'rail_execute',
         handoff_id: handoffData.rail_handoff.handoff.handoff_id,
@@ -2277,8 +2355,10 @@ async function handleStatus(projectDir) {
   }
 
   // Build model map for all agents
+  const runtimePipeline = getRuntimeAgentPipeline(projectDir, session);
+  const runtimeAgentNames = new Set(runtimePipeline.map(({ name }) => name));
   const agentModels = {};
-  for (const agentDef of AGENT_PIPELINE) {
+  for (const agentDef of runtimePipeline) {
     agentModels[agentDef.name] = resolveAgentModel(agentDef.name, projectDir);
   }
 
@@ -2309,9 +2389,9 @@ async function handleStatus(projectDir) {
     handoff_history: handoffHistory,
     deviation_count: (session.deviations || []).length,
     backlog: session.backlog || [],
-    agents: session.agents || {},
+    agents: Object.fromEntries(Object.entries(session.agents || {}).filter(([name]) => runtimeAgentNames.has(name))),
     agent_models: agentModels,
-    context_bracket: estimateContextBracket((session.completed_agents || []).length, AGENT_PIPELINE.length),
+    context_bracket: estimateContextBracket((session.completed_agents || []).length, runtimePipeline.length),
     context: {
       model:             activeModel,
       provider:          activeProvider,
@@ -2434,15 +2514,25 @@ async function handleExit(projectDir, args) {
 }
 
 async function handleProviders(projectDir) {
+  const artifact = loadRuntimeInstallationV2(projectDir);
+  if (artifact) {
+    const { session } = loadSession(projectDir);
+    const planningAgentModels = {};
+    for (const { name } of getRuntimeAgentPipeline(projectDir, session)) {
+      planningAgentModels[name] = resolveAgentModel(name, projectDir);
+    }
+    return {
+      ...describeRuntimeInstallationV2(artifact),
+      planning_agent_models: planningAgentModels,
+    };
+  }
+
   const agentModels = {};
   for (const agentDef of AGENT_PIPELINE) {
     agentModels[agentDef.name] = resolveAgentModel(agentDef.name, projectDir);
   }
 
-  const artifact = loadRuntimeInstallationV2(projectDir);
-  const enabledProviders = artifact
-    ? artifact.installation.enabled_providers.map(({ provider_id, harness_id }) => ({ provider_id, harness_id }))
-    : [...new Set(Object.values(agentModels).map(({ provider }) => provider).filter(Boolean))].map((harness_id) => ({ harness_id }));
+  const enabledProviders = [...new Set(Object.values(agentModels).map(({ provider }) => provider).filter(Boolean))].map((harness_id) => ({ harness_id }));
 
   return {
     enabled_providers: enabledProviders,

package/src/orchestrator/pipeline-manager.js

@@ -911,6 +911,8 @@ export function getPipelineProgress(pipelineState) {
   return {
     phase,
     progress: Math.round(progress),
+    completedCount: completed,
+    totalAgents,
     completedAgents,
     currentAgent,
     nextAgent,

package/src/orchestrator/runtime-installation-v2.js

@@ -10,6 +10,44 @@ const AGENT_ACTIONS = Object.freeze({
   devops: 'build',
 });
 
+export const V2_PLANNING_AGENTS = Object.freeze([
+  'greenfield-wu', 'brownfield-wu', 'brief',
+  'detail', 'architect', 'ux', 'qa-planning', 'phases', 'tasks',
+]);
+
+const RAIL_ACTIONS = new Set(['build', 'review']);
+
+export function describeRuntimeInstallationV2(artifact) {
+  const railExecutionPool = artifact.installation.enabled_providers.map((binding) => {
+    const capabilities = binding.allowed_models.flatMap((modelId) => {
+      const model = artifact.capability_snapshot.models.find((candidate) =>
+        candidate.provider_id === binding.provider_id && candidate.model_id === modelId
+      );
+      if (!model) return [];
+      const actions = model.actions.filter((action) => RAIL_ACTIONS.has(action));
+      if (actions.length === 0) return [];
+      return [{ model_id: modelId, actions, tier: model.tier || null }];
+    });
+    return {
+      provider_id: binding.provider_id,
+      harness_id: binding.harness_id,
+      models: capabilities.map(({ model_id }) => model_id),
+      capabilities,
+    };
+  }).filter(({ models }) => models.length > 0);
+
+  return Object.freeze({
+    architecture: 'chati-v2-rail',
+    enabled_providers: artifact.installation.enabled_providers.map(({ provider_id, harness_id }) => ({ provider_id, harness_id })),
+    rail_execution_pool: railExecutionPool,
+    release_lane: {
+      authority: 'explicit-human-authorization',
+      execution_authority: 'rail',
+      legacy_devops_agent: false,
+    },
+  });
+}
+
 export function loadRuntimeInstallationV2(projectDir, { clock = () => new Date() } = {}) {
   const path = join(projectDir, INSTALLATION_ARTIFACT_PATH);
   if (!existsSync(path)) return null;

package/src/orchestrator/session-manager.js

@@ -973,7 +973,7 @@ export function validateSession(projectDir) {
   }
 
   // Validate mode
-  const validModes = ['discover', 'plan', 'build', 'validate', 'deploy', 'completed'];
+  const validModes = ['discover', 'plan', 'rail', 'release', 'build', 'validate', 'deploy', 'completed'];
   if (!validModes.includes(session.mode)) {
     return {
       exists: true,