@nerviq/cli 1.20.1 → 1.21.0

package/src/harmony/advisor.js

@@ -1,333 +1,333 @@
-/**
- * H5. Strategic Project Advisor
- *
- * The strategic brain that understands a project holistically and recommends
- * which platform to use for which task type, configuration improvements,
- * and cross-platform update actions.
- *
- * Zero external dependencies - imports only from sibling/parent modules.
- */
-
-// ─── Platform Strength Matrix (canonical source: shared/capabilities.js) ─────
-
-const { PLATFORM_CAPABILITIES: PLATFORM_STRENGTHS } = require('../shared/capabilities');
-
-// ─── Task-type to platform-strength mapping ───────────────────────────────────
-
-const TASK_TYPE_PROFILES = {
-  'bug-fix': {
-    label: 'Bug Fixing',
-    requiredStrengths: { reasoning: 0.5, refactoring: 0.3, context: 0.2 },
-    description: 'Deep reasoning and code understanding for root cause analysis.',
-  },
-  'ci-review': {
-    label: 'CI / Async Review',
-    requiredStrengths: { ci: 0.4, cloudTasks: 0.3, async: 0.2, automation: 0.1 },
-    description: 'Asynchronous code review and CI-integrated workflows.',
-  },
-  'ui-work': {
-    label: 'UI Development',
-    requiredStrengths: { ide: 0.4, ui: 0.3, inline: 0.3 },
-    description: 'Interactive UI work with live preview and visual feedback.',
-  },
-  'quick-edit': {
-    label: 'Quick Edits',
-    requiredStrengths: { inline: 0.5, ide: 0.3, automation: 0.2 },
-    description: 'Small, fast inline edits and completions.',
-  },
-  'infrastructure': {
-    label: 'Infrastructure',
-    requiredStrengths: { sandbox: 0.4, cloudTasks: 0.3, reasoning: 0.3 },
-    description: 'Infrastructure, DevOps, and sandbox-heavy workflows.',
-  },
-  'refactoring': {
-    label: 'Large Refactoring',
-    requiredStrengths: { refactoring: 0.4, reasoning: 0.3, context: 0.3 },
-    description: 'Cross-file refactoring with deep codebase understanding.',
-  },
-  'code-review': {
-    label: 'Code Review',
-    requiredStrengths: { reasoning: 0.4, governance: 0.3, ci: 0.3 },
-    description: 'Thorough review with governance and standards enforcement.',
-  },
-  'greenfield': {
-    label: 'Greenfield / Scaffolding',
-    requiredStrengths: { reasoning: 0.3, automation: 0.3, ide: 0.2, refactoring: 0.2 },
-    description: 'Starting new projects or major new features from scratch.',
-  },
-  'harness-optimization': {
-    label: 'Config Optimization',
-    requiredStrengths: { governance: 0.4, automation: 0.3, reasoning: 0.3 },
-    description: 'Optimizing AI coding agent configuration and harness settings.',
-  },
-  'phased-migration': {
-    label: 'Phased Migration',
-    requiredStrengths: { reasoning: 0.4, refactoring: 0.3, architecture: 0.3 },
-    description: 'Multi-phase migrations requiring sequential execution with validation gates.',
-  },
-};
-
-// ─── Scoring ──────────────────────────────────────────────────────────────────
-
-/**
- * Score a platform against a task type profile.
- * Returns a 0-5 weighted score.
- */
-function scorePlatformForTask(platformKey, taskProfile) {
-  const strengths = PLATFORM_STRENGTHS[platformKey];
-  if (!strengths) return 0;
-
-  let score = 0;
-  for (const [dimension, weight] of Object.entries(taskProfile.requiredStrengths)) {
-    const platformScore = strengths[dimension] || 0;
-    score += platformScore * weight;
-  }
-  return Math.round(score * 100) / 100;
-}
-
-/**
- * Rank all platforms for a given task type.
- */
-function rankPlatformsForTask(taskType) {
-  const profile = TASK_TYPE_PROFILES[taskType];
-  if (!profile) return [];
-
-  const rankings = Object.keys(PLATFORM_STRENGTHS).map(platformKey => ({
-    platform: platformKey,
-    label: PLATFORM_STRENGTHS[platformKey].label,
-    score: scorePlatformForTask(platformKey, profile),
-  }));
-
-  rankings.sort((a, b) => b.score - a.score);
-  return rankings;
-}
-
-/**
- * Compute a confidence level from the score gap between #1 and #2.
- */
-function computeConfidence(rankings) {
-  if (rankings.length < 2) return 'high';
-  const gap = rankings[0].score - rankings[1].score;
-  if (gap >= 1.0) return 'high';
-  if (gap >= 0.4) return 'medium';
-  return 'low';
-}
-
-// ─── Task Routing ─────────────────────────────────────────────────────────────
-
-/**
- * Generate task routing recommendations for all known task types.
- * Optionally filter to only platforms present in platformAudits.
- */
-function generateTaskRouting(platformAudits) {
-  const availablePlatforms = platformAudits
-    ? new Set(platformAudits.map(a => a.platform))
-    : null;
-
-  const routing = [];
-
-  for (const [taskType, profile] of Object.entries(TASK_TYPE_PROFILES)) {
-    let rankings = rankPlatformsForTask(taskType);
-
-    if (availablePlatforms) {
-      rankings = rankings.filter(r => availablePlatforms.has(r.platform));
-    }
-
-    if (rankings.length === 0) continue;
-
-    const best = rankings[0];
-    routing.push({
-      taskType,
-      taskLabel: profile.label,
-      recommendedPlatform: best.platform,
-      recommendedLabel: best.label,
-      score: best.score,
-      confidence: computeConfidence(rankings),
-      reasoning: profile.description,
-      alternatives: rankings.slice(1, 3).map(r => ({
-        platform: r.platform,
-        label: r.label,
-        score: r.score,
-      })),
-    });
-  }
-
-  return routing;
-}
-
-// ─── Config Recommendations ───────────────────────────────────────────────────
-
-/**
- * Detect configuration improvement opportunities from platform audits.
- */
-function generateConfigRecommendations(canonicalModel, platformAudits) {
-  const recommendations = [];
-
-  if (!platformAudits || platformAudits.length === 0) return recommendations;
-
-  for (const auditResult of platformAudits) {
-    const platform = auditResult.platform;
-    const score = auditResult.score || 0;
-
-    // Low score platforms need attention
-    if (score < 40) {
-      recommendations.push({
-        platform,
-        recommendation: `Platform "${platform}" scores ${score}/100. Run setup to bootstrap baseline configuration.`,
-        impact: 'high',
-        category: 'bootstrap',
-      });
-    }
-
-    // Missing canonical sections
-    if (canonicalModel && canonicalModel.sections) {
-      const auditSections = new Set((auditResult.sections || []).map(s => s.key || s));
-      for (const section of canonicalModel.sections) {
-        if (!auditSections.has(section.key)) {
-          recommendations.push({
-            platform,
-            recommendation: `Missing canonical section "${section.key}" (${section.label || section.key}). Add to align with cross-platform standard.`,
-            impact: 'medium',
-            category: 'alignment',
-          });
-        }
-      }
-    }
-
-    // Permission/governance gaps
-    if (auditResult.governance) {
-      const gov = auditResult.governance;
-      if (!gov.hasPermissions) {
-        recommendations.push({
-          platform,
-          recommendation: 'No explicit permission configuration detected. Add permission boundaries for safer operation.',
-          impact: 'high',
-          category: 'security',
-        });
-      }
-      if (!gov.hasSecretProtection) {
-        recommendations.push({
-          platform,
-          recommendation: 'No secret protection hook found. Add protection to prevent credential exposure.',
-          impact: 'high',
-          category: 'security',
-        });
-      }
-    }
-  }
-
-  // Sort by impact
-  const impactOrder = { high: 0, medium: 1, low: 2 };
-  recommendations.sort((a, b) => (impactOrder[a.impact] || 2) - (impactOrder[b.impact] || 2));
-
-  return recommendations;
-}
-
-// ─── Cross-Platform Actions ───────────────────────────────────────────────────
-
-/**
- * Detect when project changes need cross-platform updates.
- * Prioritize by how many platforms are affected.
- */
-function generateCrossPlatformActions(canonicalModel, platformAudits) {
-  const actions = [];
-
-  if (!platformAudits || platformAudits.length < 2) return actions;
-
-  const platforms = platformAudits.map(a => a.platform);
-
-  // Detect instruction drift: if one platform has instructions another lacks
-  if (canonicalModel && canonicalModel.sections) {
-    const sectionCoverage = {};
-    for (const section of canonicalModel.sections) {
-      sectionCoverage[section.key] = [];
-    }
-
-    for (const auditResult of platformAudits) {
-      for (const section of (auditResult.sections || [])) {
-        const key = section.key || section;
-        if (sectionCoverage[key]) {
-          sectionCoverage[key].push(auditResult.platform);
-        }
-      }
-    }
-
-    for (const [sectionKey, coveredPlatforms] of Object.entries(sectionCoverage)) {
-      if (coveredPlatforms.length > 0 && coveredPlatforms.length < platforms.length) {
-        const missing = platforms.filter(p => !coveredPlatforms.includes(p));
-        actions.push({
-          action: `Sync section "${sectionKey}" to missing platforms`,
-          affectedPlatforms: missing,
-          sourcePlatforms: coveredPlatforms,
-          priority: missing.length >= platforms.length / 2 ? 'high' : 'medium',
-          type: 'instruction-sync',
-        });
-      }
-    }
-  }
-
-  // Detect MCP server alignment gaps
-  const mcpByPlatform = {};
-  for (const auditResult of platformAudits) {
-    mcpByPlatform[auditResult.platform] = new Set(auditResult.mcpServers || []);
-  }
-
-  const allMcpServers = new Set();
-  for (const servers of Object.values(mcpByPlatform)) {
-    for (const s of servers) allMcpServers.add(s);
-  }
-
-  for (const server of allMcpServers) {
-    const hasPlatforms = platforms.filter(p => mcpByPlatform[p] && mcpByPlatform[p].has(server));
-    const missingPlatforms = platforms.filter(p => !mcpByPlatform[p] || !mcpByPlatform[p].has(server));
-
-    if (hasPlatforms.length > 0 && missingPlatforms.length > 0) {
-      actions.push({
-        action: `Add MCP server "${server}" to platforms that lack it`,
-        affectedPlatforms: missingPlatforms,
-        sourcePlatforms: hasPlatforms,
-        priority: missingPlatforms.length > hasPlatforms.length ? 'medium' : 'low',
-        type: 'mcp-sync',
-      });
-    }
-  }
-
-  // Sort by priority then by affected platform count
-  const priorityOrder = { high: 0, medium: 1, low: 2 };
-  actions.sort((a, b) => {
-    const pDiff = (priorityOrder[a.priority] || 2) - (priorityOrder[b.priority] || 2);
-    if (pDiff !== 0) return pDiff;
-    return b.affectedPlatforms.length - a.affectedPlatforms.length;
-  });
-
-  return actions;
-}
-
-// ─── Main Entry ───────────────────────────────────────────────────────────────
-
-/**
- * Generate a complete strategic advice package.
- *
- * @param {Object|null} canonicalModel - The canonical model from canon.js
- * @param {Array} platformAudits - Array of per-platform audit results
- * @returns {Object} { taskRouting, configRecommendations, crossPlatformActions }
- */
-function generateStrategicAdvice(canonicalModel, platformAudits) {
-  return {
-    taskRouting: generateTaskRouting(platformAudits),
-    configRecommendations: generateConfigRecommendations(canonicalModel, platformAudits),
-    crossPlatformActions: generateCrossPlatformActions(canonicalModel, platformAudits),
-    generatedAt: new Date().toISOString(),
-  };
-}
-
-module.exports = {
-  generateStrategicAdvice,
-  PLATFORM_STRENGTHS,
-  TASK_TYPE_PROFILES,
-  scorePlatformForTask,
-  rankPlatformsForTask,
-  generateTaskRouting,
-  generateConfigRecommendations,
-  generateCrossPlatformActions,
-};
+/**
+ * H5. Strategic Project Advisor
+ *
+ * The strategic brain that understands a project holistically and recommends
+ * which platform to use for which task type, configuration improvements,
+ * and cross-platform update actions.
+ *
+ * Zero external dependencies - imports only from sibling/parent modules.
+ */
+
+// ─── Platform Strength Matrix (canonical source: shared/capabilities.js) ─────
+
+const { PLATFORM_CAPABILITIES: PLATFORM_STRENGTHS } = require('../shared/capabilities');
+
+// ─── Task-type to platform-strength mapping ───────────────────────────────────
+
+const TASK_TYPE_PROFILES = {
+  'bug-fix': {
+    label: 'Bug Fixing',
+    requiredStrengths: { reasoning: 0.5, refactoring: 0.3, context: 0.2 },
+    description: 'Deep reasoning and code understanding for root cause analysis.',
+  },
+  'ci-review': {
+    label: 'CI / Async Review',
+    requiredStrengths: { ci: 0.4, cloudTasks: 0.3, async: 0.2, automation: 0.1 },
+    description: 'Asynchronous code review and CI-integrated workflows.',
+  },
+  'ui-work': {
+    label: 'UI Development',
+    requiredStrengths: { ide: 0.4, ui: 0.3, inline: 0.3 },
+    description: 'Interactive UI work with live preview and visual feedback.',
+  },
+  'quick-edit': {
+    label: 'Quick Edits',
+    requiredStrengths: { inline: 0.5, ide: 0.3, automation: 0.2 },
+    description: 'Small, fast inline edits and completions.',
+  },
+  'infrastructure': {
+    label: 'Infrastructure',
+    requiredStrengths: { sandbox: 0.4, cloudTasks: 0.3, reasoning: 0.3 },
+    description: 'Infrastructure, DevOps, and sandbox-heavy workflows.',
+  },
+  'refactoring': {
+    label: 'Large Refactoring',
+    requiredStrengths: { refactoring: 0.4, reasoning: 0.3, context: 0.3 },
+    description: 'Cross-file refactoring with deep codebase understanding.',
+  },
+  'code-review': {
+    label: 'Code Review',
+    requiredStrengths: { reasoning: 0.4, governance: 0.3, ci: 0.3 },
+    description: 'Thorough review with governance and standards enforcement.',
+  },
+  'greenfield': {
+    label: 'Greenfield / Scaffolding',
+    requiredStrengths: { reasoning: 0.3, automation: 0.3, ide: 0.2, refactoring: 0.2 },
+    description: 'Starting new projects or major new features from scratch.',
+  },
+  'harness-optimization': {
+    label: 'Config Optimization',
+    requiredStrengths: { governance: 0.4, automation: 0.3, reasoning: 0.3 },
+    description: 'Optimizing AI coding agent configuration and harness settings.',
+  },
+  'phased-migration': {
+    label: 'Phased Migration',
+    requiredStrengths: { reasoning: 0.4, refactoring: 0.3, architecture: 0.3 },
+    description: 'Multi-phase migrations requiring sequential execution with validation gates.',
+  },
+};
+
+// ─── Scoring ──────────────────────────────────────────────────────────────────
+
+/**
+ * Score a platform against a task type profile.
+ * Returns a 0-5 weighted score.
+ */
+function scorePlatformForTask(platformKey, taskProfile) {
+  const strengths = PLATFORM_STRENGTHS[platformKey];
+  if (!strengths) return 0;
+
+  let score = 0;
+  for (const [dimension, weight] of Object.entries(taskProfile.requiredStrengths)) {
+    const platformScore = strengths[dimension] || 0;
+    score += platformScore * weight;
+  }
+  return Math.round(score * 100) / 100;
+}
+
+/**
+ * Rank all platforms for a given task type.
+ */
+function rankPlatformsForTask(taskType) {
+  const profile = TASK_TYPE_PROFILES[taskType];
+  if (!profile) return [];
+
+  const rankings = Object.keys(PLATFORM_STRENGTHS).map(platformKey => ({
+    platform: platformKey,
+    label: PLATFORM_STRENGTHS[platformKey].label,
+    score: scorePlatformForTask(platformKey, profile),
+  }));
+
+  rankings.sort((a, b) => b.score - a.score);
+  return rankings;
+}
+
+/**
+ * Compute a confidence level from the score gap between #1 and #2.
+ */
+function computeConfidence(rankings) {
+  if (rankings.length < 2) return 'high';
+  const gap = rankings[0].score - rankings[1].score;
+  if (gap >= 1.0) return 'high';
+  if (gap >= 0.4) return 'medium';
+  return 'low';
+}
+
+// ─── Task Routing ─────────────────────────────────────────────────────────────
+
+/**
+ * Generate task routing recommendations for all known task types.
+ * Optionally filter to only platforms present in platformAudits.
+ */
+function generateTaskRouting(platformAudits) {
+  const availablePlatforms = platformAudits
+    ? new Set(platformAudits.map(a => a.platform))
+    : null;
+
+  const routing = [];
+
+  for (const [taskType, profile] of Object.entries(TASK_TYPE_PROFILES)) {
+    let rankings = rankPlatformsForTask(taskType);
+
+    if (availablePlatforms) {
+      rankings = rankings.filter(r => availablePlatforms.has(r.platform));
+    }
+
+    if (rankings.length === 0) continue;
+
+    const best = rankings[0];
+    routing.push({
+      taskType,
+      taskLabel: profile.label,
+      recommendedPlatform: best.platform,
+      recommendedLabel: best.label,
+      score: best.score,
+      confidence: computeConfidence(rankings),
+      reasoning: profile.description,
+      alternatives: rankings.slice(1, 3).map(r => ({
+        platform: r.platform,
+        label: r.label,
+        score: r.score,
+      })),
+    });
+  }
+
+  return routing;
+}
+
+// ─── Config Recommendations ───────────────────────────────────────────────────
+
+/**
+ * Detect configuration improvement opportunities from platform audits.
+ */
+function generateConfigRecommendations(canonicalModel, platformAudits) {
+  const recommendations = [];
+
+  if (!platformAudits || platformAudits.length === 0) return recommendations;
+
+  for (const auditResult of platformAudits) {
+    const platform = auditResult.platform;
+    const score = auditResult.score || 0;
+
+    // Low score platforms need attention
+    if (score < 40) {
+      recommendations.push({
+        platform,
+        recommendation: `Platform "${platform}" scores ${score}/100. Run setup to bootstrap baseline configuration.`,
+        impact: 'high',
+        category: 'bootstrap',
+      });
+    }
+
+    // Missing canonical sections
+    if (canonicalModel && canonicalModel.sections) {
+      const auditSections = new Set((auditResult.sections || []).map(s => s.key || s));
+      for (const section of canonicalModel.sections) {
+        if (!auditSections.has(section.key)) {
+          recommendations.push({
+            platform,
+            recommendation: `Missing canonical section "${section.key}" (${section.label || section.key}). Add to align with cross-platform standard.`,
+            impact: 'medium',
+            category: 'alignment',
+          });
+        }
+      }
+    }
+
+    // Permission/governance gaps
+    if (auditResult.governance) {
+      const gov = auditResult.governance;
+      if (!gov.hasPermissions) {
+        recommendations.push({
+          platform,
+          recommendation: 'No explicit permission configuration detected. Add permission boundaries for safer operation.',
+          impact: 'high',
+          category: 'security',
+        });
+      }
+      if (!gov.hasSecretProtection) {
+        recommendations.push({
+          platform,
+          recommendation: 'No secret protection hook found. Add protection to prevent credential exposure.',
+          impact: 'high',
+          category: 'security',
+        });
+      }
+    }
+  }
+
+  // Sort by impact
+  const impactOrder = { high: 0, medium: 1, low: 2 };
+  recommendations.sort((a, b) => (impactOrder[a.impact] || 2) - (impactOrder[b.impact] || 2));
+
+  return recommendations;
+}
+
+// ─── Cross-Platform Actions ───────────────────────────────────────────────────
+
+/**
+ * Detect when project changes need cross-platform updates.
+ * Prioritize by how many platforms are affected.
+ */
+function generateCrossPlatformActions(canonicalModel, platformAudits) {
+  const actions = [];
+
+  if (!platformAudits || platformAudits.length < 2) return actions;
+
+  const platforms = platformAudits.map(a => a.platform);
+
+  // Detect instruction drift: if one platform has instructions another lacks
+  if (canonicalModel && canonicalModel.sections) {
+    const sectionCoverage = {};
+    for (const section of canonicalModel.sections) {
+      sectionCoverage[section.key] = [];
+    }
+
+    for (const auditResult of platformAudits) {
+      for (const section of (auditResult.sections || [])) {
+        const key = section.key || section;
+        if (sectionCoverage[key]) {
+          sectionCoverage[key].push(auditResult.platform);
+        }
+      }
+    }
+
+    for (const [sectionKey, coveredPlatforms] of Object.entries(sectionCoverage)) {
+      if (coveredPlatforms.length > 0 && coveredPlatforms.length < platforms.length) {
+        const missing = platforms.filter(p => !coveredPlatforms.includes(p));
+        actions.push({
+          action: `Sync section "${sectionKey}" to missing platforms`,
+          affectedPlatforms: missing,
+          sourcePlatforms: coveredPlatforms,
+          priority: missing.length >= platforms.length / 2 ? 'high' : 'medium',
+          type: 'instruction-sync',
+        });
+      }
+    }
+  }
+
+  // Detect MCP server alignment gaps
+  const mcpByPlatform = {};
+  for (const auditResult of platformAudits) {
+    mcpByPlatform[auditResult.platform] = new Set(auditResult.mcpServers || []);
+  }
+
+  const allMcpServers = new Set();
+  for (const servers of Object.values(mcpByPlatform)) {
+    for (const s of servers) allMcpServers.add(s);
+  }
+
+  for (const server of allMcpServers) {
+    const hasPlatforms = platforms.filter(p => mcpByPlatform[p] && mcpByPlatform[p].has(server));
+    const missingPlatforms = platforms.filter(p => !mcpByPlatform[p] || !mcpByPlatform[p].has(server));
+
+    if (hasPlatforms.length > 0 && missingPlatforms.length > 0) {
+      actions.push({
+        action: `Add MCP server "${server}" to platforms that lack it`,
+        affectedPlatforms: missingPlatforms,
+        sourcePlatforms: hasPlatforms,
+        priority: missingPlatforms.length > hasPlatforms.length ? 'medium' : 'low',
+        type: 'mcp-sync',
+      });
+    }
+  }
+
+  // Sort by priority then by affected platform count
+  const priorityOrder = { high: 0, medium: 1, low: 2 };
+  actions.sort((a, b) => {
+    const pDiff = (priorityOrder[a.priority] || 2) - (priorityOrder[b.priority] || 2);
+    if (pDiff !== 0) return pDiff;
+    return b.affectedPlatforms.length - a.affectedPlatforms.length;
+  });
+
+  return actions;
+}
+
+// ─── Main Entry ───────────────────────────────────────────────────────────────
+
+/**
+ * Generate a complete strategic advice package.
+ *
+ * @param {Object|null} canonicalModel - The canonical model from canon.js
+ * @param {Array} platformAudits - Array of per-platform audit results
+ * @returns {Object} { taskRouting, configRecommendations, crossPlatformActions }
+ */
+function generateStrategicAdvice(canonicalModel, platformAudits) {
+  return {
+    taskRouting: generateTaskRouting(platformAudits),
+    configRecommendations: generateConfigRecommendations(canonicalModel, platformAudits),
+    crossPlatformActions: generateCrossPlatformActions(canonicalModel, platformAudits),
+    generatedAt: new Date().toISOString(),
+  };
+}
+
+module.exports = {
+  generateStrategicAdvice,
+  PLATFORM_STRENGTHS,
+  TASK_TYPE_PROFILES,
+  scorePlatformForTask,
+  rankPlatformsForTask,
+  generateTaskRouting,
+  generateConfigRecommendations,
+  generateCrossPlatformActions,
+};