chati-dev 2.1.2 → 3.0.1

package/framework/tasks/brownfield-wu-migration-plan.md

@@ -67,18 +67,18 @@ Present options and recommend:
 
 ### 3. Define Migration Phases
 Break migration into phases (typically 4-8 phases):
-- **Phase 0: Preparation**
+- **Phase 1: Preparation**
   - Set up monitoring and observability
   - Increase test coverage for critical paths
   - Document current behavior
   - Fix critical security issues
   - Set up feature flags
-- **Phase 1: Foundation**
+- **Phase 2: Foundation**
   - Update dependencies to secure versions
   - Standardize tooling (linting, formatting, testing)
   - Resolve circular dependencies
   - Set up CI/CD if missing
-- **Phase 2-N: Incremental Changes**
+- **Phase 3-N: Incremental Changes**
   - Each phase targets a specific subsystem or concern
   - Maintain working system at end of each phase
   - Deploy and validate before next phase
@@ -152,7 +152,7 @@ For each phase, document:
 - Add 20-30% buffer for unknowns
 - Calculate calendar time based on:
   - Team size
-  - Velocity (story points per sprint)
+  - Velocity (estimated hours per phase)
   - Availability (not 100% on migration)
   - Dependencies and waiting time
 - Provide optimistic, realistic, and pessimistic timelines

package/framework/tasks/brownfield-wu-risk-assess.md

@@ -159,7 +159,7 @@ For each identified risk:
 ### 9. Prioritize Risks
 Sort risks by risk score descending:
 - **Critical Priority (Score 70-100)**: Address immediately
-- **High Priority (Score 40-69)**: Address in next sprint
+- **High Priority (Score 40-69)**: Address in next phase
 - **Medium Priority (Score 20-39)**: Address within 3 months
 - **Low Priority (Score 1-19)**: Monitor, address when convenient
 
@@ -419,6 +419,6 @@ total_remediation_effort:
 next_steps:
   - Address 3 critical security risks immediately
   - Create detailed remediation plan for high priority risks
-  - Schedule technical debt reduction sprints
+  - Schedule technical debt reduction phases
   - Set up monitoring to track new risks
 ```

package/framework/tasks/detail-expand-prd.md

@@ -10,7 +10,7 @@ handoff_to: detail-nfr-extraction
 autonomous_gate: true
 criteria:
   - All functional requirements expanded with details
-  - User stories written in Given-When-Then format
+  - Requirements written in Given-When-Then format
   - Data models identified
   - API endpoints outlined
 ---
@@ -30,7 +30,7 @@ Read brief.yaml and extract all requirements and context.
 ### 2. Expand Each Functional Requirement
 For each FR, add:
 - **Description**: Detailed explanation of what it does
-- **User Story**: As a [persona], I want [action] so that [benefit]
+- **Requirement**: As a [persona], I want [action] so that [benefit]
 - **Acceptance Criteria**: Given-When-Then scenarios
 - **Priority**: P0 (must-have), P1 (should-have), P2 (nice-to-have)
 - **Estimated Complexity**: Small (1-3 days), Medium (4-7 days), Large (8+ days)

package/framework/tasks/greenfield-wu-report.md

@@ -98,7 +98,7 @@ Provide ordered list of immediate actions:
 ### 9. Add Recommendations
 Include actionable recommendations:
 - **Immediate**: Actions to take before writing code
-- **Short-term**: Set up during first sprint (CI/CD, testing)
+- **Short-term**: Set up during Phase 1 (CI/CD, testing)
 - **Medium-term**: Add as project matures (monitoring, documentation site)
 - **Long-term**: Consider for scale (microservices, caching layers)
 

package/framework/tasks/orchestrator-handoff.md

@@ -175,7 +175,7 @@ handoff_document:
 
   summary: |
     Completed detailed requirements analysis for HealthCare Portal.
-    Identified 12 user stories across 4 modules (Authentication, Patient Records,
+    Identified 12 requirements across 4 modules (Authentication, Patient Records,
     Appointments, Billing). Mapped 3 external integrations (Payment Gateway,
     Insurance API, Lab Results). Defined data model with 8 core entities.
 
@@ -186,7 +186,7 @@ handoff_document:
       size_kb: 45
     - path: ".chati/artifacts/requirements/user-stories.yaml"
       type: "structured-data"
-      description: "12 user stories with acceptance criteria"
+      description: "12 requirements with acceptance criteria"
       size_kb: 8
     - path: ".chati/artifacts/requirements/integrations.yaml"
       type: "structured-data"

package/framework/tasks/orchestrator-resume.md

@@ -173,7 +173,7 @@ session_state:
     file: ".chati/artifacts/handoffs/2026-02-13-detail.md"
     key_outputs:
       - "Functional requirements document"
-      - "12 user stories identified"
+      - "12 requirements identified"
       - "3 external integrations mapped"
   context_summary:
     decisions_made: 8

package/framework/tasks/qa-impl-performance-test.md

@@ -461,7 +461,7 @@ assessment:
 
   recommendations:
     - Current performance acceptable for release
-    - Consider YAML parser alternatives in future optimization sprint
+    - Consider YAML parser alternatives in future optimization phase
     - Monitor agent execution time in production, optimize if user complaints
 
 baseline_update:
@@ -477,7 +477,7 @@ next_steps:
   - Update performance baseline with improved metrics
   - Proceed to qa-impl-verdict (performance gate PASSED)
   - Monitor performance in production after release
-  - Schedule optimization sprint for YAML parsing if user impact observed
+  - Schedule optimization phase for YAML parsing if user impact observed
 
 handoff:
   to: qa-impl-verdict

package/framework/tasks/qa-impl-sast-scan.md

@@ -81,9 +81,9 @@ Run static analysis security testing (SAST) to identify security vulnerabilities
 
 9. **Analyze Trends (if historical data available)**
    - Compare findings with previous scan (if available)
-   - Track new vulnerabilities introduced in this sprint
+   - Track new vulnerabilities introduced in this phase
    - Track resolved vulnerabilities since last scan
-   - Calculate vulnerability velocity (new - resolved per sprint)
+   - Calculate vulnerability velocity (new - resolved per phase)
 
 10. **Generate Remediation Guidance**
     - For each CRITICAL/HIGH finding:
@@ -330,7 +330,7 @@ trends:
   previous_scan_date: YYYY-MM-DD (or null if first scan)
   new_vulnerabilities: 2
   resolved_vulnerabilities: 0
-  vulnerability_velocity: +2 per sprint
+  vulnerability_velocity: +2 per phase
 
 remediation_roadmap:
   priority_1_blocking:
@@ -378,19 +378,19 @@ assessment:
     - Fix both HIGH severity exploitable vulnerabilities
     - Upgrade yaml package to 2.3.4+
     - Re-run SAST scan to verify fixes
-    - Medium and low findings can be deferred to next sprint
+    - Medium and low findings can be deferred to next phase
 
   recommendations:
     - Priority 1: Fix SAST-001 with path validation (2 hours)
     - Priority 2: Fix SAST-002 with execFile (1 hour)
     - Priority 3: Upgrade yaml dependency (30 minutes)
-    - Schedule code quality fixes for next sprint
+    - Schedule code quality fixes for next phase
 
 next_steps:
   - Return to dev agent for HIGH vulnerability remediation
   - Re-run qa-impl-sast-scan after fixes
   - If rerun passes (0 HIGH exploitable), proceed to qa-impl-regression-check
-  - Track LOW/MEDIUM findings in backlog for future sprints
+  - Track LOW/MEDIUM findings in backlog for future phases
 
 handoff:
   to: dev

package/framework/tasks/qa-planning-consolidate.md

@@ -303,7 +303,7 @@ session_update:
 
 next_steps:
   - Orchestrator triggers dev agent (autonomous transition)
-  - Dev agent begins Phase 0 (Installer) implementation
+  - Dev agent begins Phase 1 (Installer) implementation
   - QA-Implementation agent on standby for build phase
   - Revisit minor gaps during Phase 1 review
 ```

package/framework/tasks/qa-planning-coverage-plan.md

@@ -78,7 +78,7 @@ Create a detailed test coverage plan with specific targets for each module, comp
 
 8. **Establish Coverage Baselines**
    - **Initial Baseline**: Measure current coverage (if brownfield) or set 0% (if greenfield)
-   - **Sprint Targets**: Define incremental improvement goals (e.g., +5% per sprint)
+   - **Phase Targets**: Define incremental improvement goals (e.g., +5% per phase)
    - **Release Targets**: Set minimum coverage for each release milestone
    - **Maintenance Target**: Steady-state coverage after initial development (e.g., 75%)
 
@@ -95,10 +95,10 @@ Create a detailed test coverage plan with specific targets for each module, comp
     - **Review Cadence**: When exception is revisited (quarterly)
 
 11. **Create Coverage Improvement Roadmap**
-    - **Phase 0 (Installer)**: 70% target (focus on CLI and validation logic)
-    - **Phase 1-3 (Core Agents)**: 75% target (add agent task execution tests)
-    - **Phase 4-5 (Quality/Build)**: 80% target (comprehensive integration tests)
-    - **Phase 6-7 (Deploy/Validate)**: 85% target (full pipeline coverage)
+    - **Phase 1 (Installer)**: 70% target (focus on CLI and validation logic)
+    - **Phase 2-4 (Core Agents)**: 75% target (add agent task execution tests)
+    - **Phase 5-6 (Quality/Build)**: 80% target (comprehensive integration tests)
+    - **Phase 7-8 (Deploy/Validate)**: 85% target (full pipeline coverage)
 
 12. **Compile Coverage Plan Document**
     - List all modules with specific targets
@@ -108,7 +108,7 @@ Create a detailed test coverage plan with specific targets for each module, comp
 
 ## Decision Points
 
-- **Target Adjustment for Legacy Code**: If brownfield project has existing low coverage, set realistic incremental targets rather than immediate 80% requirement. Example: 40% current → 50% sprint 1 → 60% sprint 2 → 70% sprint 3.
+- **Target Adjustment for Legacy Code**: If brownfield project has existing low coverage, set realistic incremental targets rather than immediate 80% requirement. Example: 40% current → 50% Phase 1 → 60% Phase 2 → 70% Phase 3.
 
 - **Integration Test Coverage**: For external dependencies (MCPs, file system), decide between mocking (higher coverage, less realistic) vs real integration (lower coverage, more realistic). Recommend hybrid: unit tests with mocks, separate integration tests with real dependencies.
 
@@ -287,9 +287,9 @@ exceptions:
     review_date: quarterly
 
 baselines:
-  sprint_1: 70%
-  sprint_2: 75%
-  sprint_3: 80%
+  phase_1: 70%
+  phase_2: 75%
+  phase_3: 80%
   release_1.0: 75%
   maintenance: 75%
 

package/framework/tasks/qa-planning-gate-define.md

@@ -73,7 +73,7 @@ Define quality gates and enforcement thresholds for each stage of the chati.dev
 7. **Establish Threshold Levels**
    - **BLOCKER**: Must fix immediately, blocks all progress (test failures, critical vulnerabilities)
    - **CRITICAL**: Must fix before merge/release (high vulnerabilities, major regressions)
-   - **MAJOR**: Should fix in current sprint (moderate issues, technical debt)
+   - **MAJOR**: Should fix in current phase (moderate issues, technical debt)
    - **MINOR**: Can defer to backlog (low priority issues, enhancements)
 
 8. **Define Gate Enforcement Mechanisms**
@@ -98,7 +98,7 @@ Define quality gates and enforcement thresholds for each stage of the chati.dev
     - **Gate Pass Rate**: Track percentage of clean gate passes
     - **Bypass Frequency**: Monitor override usage patterns
     - **Time in Gate**: Measure delay caused by gate failures
-    - **Trend Analysis**: Track improvement/degradation over sprints
+    - **Trend Analysis**: Track improvement/degradation over phases
 
 12. **Compile Quality Gates Document**
     - Structure gates by pipeline stage
@@ -127,7 +127,7 @@ Define quality gates and enforcement thresholds for each stage of the chati.dev
 
 **Tool Integration Failures**
 - If CI/CD pipeline cannot enforce a gate automatically, define manual checklist
-- Flag for devops agent to implement automation in future sprint
+- Flag for devops agent to implement automation in future phase
 - Ensure interim manual process is documented
 
 ## Output Format
@@ -253,14 +253,14 @@ severity_levels:
     resolution_time: same day
 
   MAJOR:
-    description: Fix in current sprint
+    description: Fix in current phase
     examples: [moderate issues, technical debt, performance degradation]
-    resolution_time: within sprint
+    resolution_time: within phase
 
   MINOR:
     description: Backlog eligible
     examples: [low priority issues, enhancements, cosmetic fixes]
-    resolution_time: next sprint or later
+    resolution_time: next phase or later
 
 bypass_procedures:
   deviation_protocol:

package/framework/tasks/qa-planning-risk-matrix.md

@@ -607,7 +607,7 @@ risk_evolution:
 
   mid_phase:
     focus: [regression risks, performance risks, error handling]
-    reassessment: per sprint
+    reassessment: per phase
 
   late_phase:
     focus: [upgrade risks, compatibility risks, edge cases]

package/framework/tasks/tasks-consolidate.md

@@ -27,8 +27,8 @@ Merge: tasks-list.yaml, estimates.yaml, task-criteria.yaml
 Organize tasks hierarchically:
 - Phase → Feature → Tasks
 
-### 3. Add Sprint Planning Hints
-Suggest which tasks could be grouped into sprints.
+### 3. Add Phase Planning Hints
+Suggest which tasks could be grouped into phases.
 
 ### 4. Generate Final Task List
 Complete, ready for development.
@@ -40,12 +40,12 @@ timestamp: 2026-02-13T18:00:00Z
 total_tasks: 67
 total_estimated_hours: 480
 [All task details consolidated]
-sprint_suggestions:
-  sprint_1:
+phase_suggestions:
+  phase_1:
     focus: Database schema and authentication backend
     tasks: [TASK-001, TASK-002, TASK-005, TASK-006]
     estimated_hours: 60
-  sprint_2:
+  phase_2:
     focus: Authentication UI and post backend
     tasks: [TASK-003, TASK-004, TASK-010, TASK-011]
     estimated_hours: 72

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "chati-dev",
-  "version": "2.1.2",
+  "version": "3.0.1",
   "description": "AI-Powered Multi-Agent Orchestration System — Structured vibe coding for Full Stack Development",
   "type": "module",
   "bin": {
@@ -38,6 +38,10 @@
     "multi-agent",
     "cli",
     "claude",
+    "gemini",
+    "codex",
+    "copilot",
+    "multi-llm",
     "orchestration",
     "fullstack"
   ],

package/scripts/changelog-generator.js

@@ -7,7 +7,7 @@
  * @module scripts/changelog-generator
  */
 
-import { execSync } from 'node:child_process';
+import { execFileSync } from 'node:child_process';
 
 /**
  * @typedef {Object} ParsedCommit
@@ -84,7 +84,7 @@ export function getGitCommits(since = null, cwd = process.cwd()) {
 
   let output;
   try {
-    output = execSync(`git log ${range} --pretty=format:"${format}"`, {
+    output = execFileSync('git', ['log', range, `--pretty=format:${format}`], {
       cwd,
       encoding: 'utf-8',
       stdio: ['pipe', 'pipe', 'pipe'],

package/scripts/ide-sync.js

@@ -49,13 +49,13 @@ export const IDE_CONFIGS = {
     description: 'AntiGravity',
   },
   'gemini-cli': {
-    configDir: '.gemini/',
-    extension: '.md',
-    format: 'markdown',
+    commandDir: '.gemini/commands/',
+    extension: '.toml',
+    format: 'toml',
     description: 'Gemini CLI',
   },
   'github-copilot': {
-    instructionsDir: '.github/copilot/',
+    agentsDir: '.github/agents/',
     extension: '.md',
     format: 'markdown',
     description: 'GitHub Copilot',
@@ -154,6 +154,11 @@ export function buildIDEContent(agentContent, format, agentName = '') {
       return lines.join('\n');
     }
 
+    case 'toml': {
+      // Gemini CLI uses TOML format for commands
+      return `description = "chati.dev agent — ${agentName}"\n\nprompt = """\n${agentContent}\n"""\n`;
+    }
+
     case 'markdown':
     default:
       return agentContent;

package/scripts/pr-review.js

@@ -7,7 +7,7 @@
  * @module scripts/pr-review
  */
 
-import { execSync } from 'node:child_process';
+import { execFileSync } from 'node:child_process';
 import { extname, dirname } from 'node:path';
 
 /**
@@ -64,13 +64,13 @@ export function getSensitivePaths() {
 
 /**
  * Executes a git command and returns the output.
- * @param {string} cmd
+ * @param {string[]} args - Git subcommand and arguments
  * @param {string} cwd
  * @returns {string}
  */
-function git(cmd, cwd) {
+function git(args, cwd) {
   try {
-    return execSync(`git ${cmd}`, {
+    return execFileSync('git', args, {
       cwd,
       encoding: 'utf-8',
       stdio: ['pipe', 'pipe', 'pipe'],
@@ -89,7 +89,7 @@ function git(cmd, cwd) {
  * @returns {ChangedFiles}
  */
 export function getChangedFiles(base, head, cwd = process.cwd()) {
-  const output = git(`diff --name-status ${base}...${head}`, cwd);
+  const output = git(['diff', '--name-status', `${base}...${head}`], cwd);
   if (!output) {
     return { added: [], modified: [], deleted: [], renamed: [] };
   }

package/scripts/semantic-lint.js

@@ -185,7 +185,7 @@ function checkSchemaExistence(frameworkDir, results) {
 }
 
 /**
- * Check constitution has >= 17 articles.
+ * Check constitution has >= 19 articles.
  */
 function checkConstitution(frameworkDir, results) {
   results.checks++;
@@ -199,11 +199,11 @@ function checkConstitution(frameworkDir, results) {
   const content = readFileSync(constitutionPath, 'utf-8');
   const articleCount = (content.match(/^## Article/gm) || []).length;
 
-  if (articleCount >= 17) {
+  if (articleCount >= 19) {
     results.passed++;
     results.details.push(`Constitution: ${articleCount} articles found`);
   } else {
-    results.errors.push(`Constitution has only ${articleCount} articles (expected >= 17)`);
+    results.errors.push(`Constitution has only ${articleCount} articles (expected >= 19)`);
   }
 }
 

package/src/autonomy/build-loop.js

@@ -0,0 +1,194 @@
+/**
+ * @fileoverview Autonomous build loop (Ralph Wiggum v2).
+ *
+ * Executes tasks autonomously with checkpoint-based state management,
+ * retry logic, and quality gate integration.
+ *
+ * Named "Ralph Wiggum" internally — the autonomous execution mode
+ * that iterates until all tasks are complete or escalation is needed.
+ *
+ * Constitution Article XVII — Execution Mode Governance.
+ */
+
+import {
+  createBuildState,
+  loadBuildState,
+  saveBuildState,
+  startBuild,
+  completeBuild,
+  failBuild,
+  updateCheckpoint,
+  getNextPendingTask,
+  isTaskExhausted,
+  isTimedOut,
+  getProgress,
+  CheckpointStatus,
+  BuildStatus,
+} from './build-state.js';
+
+// ---------------------------------------------------------------------------
+// Build Loop
+// ---------------------------------------------------------------------------
+
+/**
+ * @typedef {object} BuildLoopConfig
+ * @property {string} projectDir - Project root directory
+ * @property {string[]} taskIds - Task IDs to execute
+ * @property {function(string): Promise<{success: boolean, output: string}>} executor - Task execution function
+ * @property {function(object): void} [onProgress] - Progress callback
+ * @property {boolean} [resume=false] - Whether to resume from existing state
+ */
+
+/**
+ * @typedef {object} BuildLoopResult
+ * @property {string} status - Final build status
+ * @property {number} completed - Tasks completed
+ * @property {number} failed - Tasks failed
+ * @property {number} totalAttempts - Total execution attempts
+ * @property {string} duration - Human-readable duration
+ */
+
+/**
+ * Run the autonomous build loop.
+ *
+ * Loop logic:
+ * 1. Load or create build state
+ * 2. Get next pending task
+ * 3. Execute task
+ * 4. Save checkpoint
+ * 5. If task failed and not exhausted, retry
+ * 6. If task exhausted, mark as failed and continue
+ * 7. Repeat until all tasks complete or global timeout
+ *
+ * @param {BuildLoopConfig} config
+ * @returns {Promise<BuildLoopResult>}
+ */
+export async function runBuildLoop(config) {
+  const { projectDir, taskIds, executor, onProgress, resume = false } = config;
+
+  // Load or create state
+  let state = resume ? loadBuildState(projectDir) : null;
+
+  if (!state || state.status === BuildStatus.COMPLETED || state.status === BuildStatus.ABANDONED) {
+    state = createBuildState(taskIds);
+  }
+
+  state = startBuild(state);
+  saveBuildState(projectDir, state);
+
+  const startTime = Date.now();
+
+  // Main loop
+  while (true) {
+    // Check global timeout
+    if (isTimedOut(state)) {
+      state = failBuild(state, 'Global timeout exceeded');
+      saveBuildState(projectDir, state);
+      break;
+    }
+
+    // Get next task
+    const checkpoint = getNextPendingTask(state);
+    if (!checkpoint) {
+      // All tasks processed
+      const hasFailures = state.checkpoints.some((c) => c.status === CheckpointStatus.FAILED);
+      if (hasFailures) {
+        state = failBuild(state, 'Some tasks failed');
+      } else {
+        state = completeBuild(state);
+      }
+      saveBuildState(projectDir, state);
+      break;
+    }
+
+    // Check if task is exhausted
+    if (isTaskExhausted(checkpoint)) {
+      state = updateCheckpoint(state, checkpoint.taskId, {
+        status: CheckpointStatus.FAILED,
+        error: `Exceeded max iterations (${checkpoint.attempts})`,
+      });
+      saveBuildState(projectDir, state);
+
+      if (onProgress) {
+        onProgress({ type: 'task_exhausted', taskId: checkpoint.taskId, attempts: checkpoint.attempts });
+      }
+      continue;
+    }
+
+    // Mark task as in progress
+    state = updateCheckpoint(state, checkpoint.taskId, {
+      status: CheckpointStatus.IN_PROGRESS,
+      attempts: checkpoint.attempts + 1,
+      lastAttempt: new Date().toISOString(),
+    });
+    saveBuildState(projectDir, state);
+
+    if (onProgress) {
+      const progress = getProgress(state);
+      onProgress({ type: 'task_started', taskId: checkpoint.taskId, attempt: checkpoint.attempts + 1, progress });
+    }
+
+    // Execute task
+    try {
+      const result = await executor(checkpoint.taskId);
+
+      if (result.success) {
+        state = updateCheckpoint(state, checkpoint.taskId, {
+          status: CheckpointStatus.COMPLETED,
+          output: result.output?.slice(0, 1000) || 'Completed',
+          error: null,
+        });
+
+        if (onProgress) {
+          onProgress({ type: 'task_completed', taskId: checkpoint.taskId });
+        }
+      } else {
+        state = updateCheckpoint(state, checkpoint.taskId, {
+          status: CheckpointStatus.IN_PROGRESS, // Will retry
+          error: result.output?.slice(0, 500) || 'Task failed',
+        });
+
+        if (onProgress) {
+          onProgress({ type: 'task_failed', taskId: checkpoint.taskId, attempt: checkpoint.attempts + 1, error: result.output });
+        }
+      }
+    } catch (err) {
+      state = updateCheckpoint(state, checkpoint.taskId, {
+        status: CheckpointStatus.IN_PROGRESS, // Will retry
+        error: err.message?.slice(0, 500) || 'Execution error',
+      });
+    }
+
+    saveBuildState(projectDir, state);
+  }
+
+  const duration = Date.now() - startTime;
+  const progress = getProgress(state);
+
+  return {
+    status: state.status,
+    completed: progress.completed,
+    failed: progress.failed,
+    totalAttempts: state.totalAttempts,
+    duration: `${Math.round(duration / 1000)}s`,
+  };
+}
+
+/**
+ * Get current build loop status without executing.
+ *
+ * @param {string} projectDir
+ * @returns {object|null}
+ */
+export function getBuildStatus(projectDir) {
+  const state = loadBuildState(projectDir);
+  if (!state) return null;
+
+  return {
+    sessionId: state.sessionId,
+    status: state.status,
+    progress: getProgress(state),
+    startedAt: state.startedAt,
+    lastCheckpoint: state.lastCheckpoint,
+  };
+}