tractatus/scripts/hook-validators/validate-file-write.js

#!/usr/bin/env node

/**
 * Hook Validator: File Write
 *
 * Runs BEFORE Write tool execution to enforce governance requirements.
 * This is architectural enforcement - AI cannot bypass this check.
 *
 * Checks:
 * 1. Pre-action validation (CSP, file type restrictions)
 * 2. Overwrite without read check (file exists but not read)
 * 3. CrossReferenceValidator (instruction conflicts)
 * 4. BoundaryEnforcer (values decisions)
 *
 * Exit codes:
 * 0 = PASS (allow write)
 * 1 = FAIL (block write)
 *
 * Copyright 2025 Tractatus Project
 * Licensed under Apache License 2.0
 */

const fs = require('fs');
const path = require('path');
const { execSync } = require('child_process');

// Import CrossReferenceValidator
const CrossReferenceValidator = require('../framework-components/CrossReferenceValidator.js');

// Hooks receive input via stdin as JSON
let HOOK_INPUT = null;
let FILE_PATH = null;

const SESSION_STATE_PATH = path.join(__dirname, '../../.claude/session-state.json');
const INSTRUCTION_HISTORY_PATH = path.join(__dirname, '../../.claude/instruction-history.json');

/**
 * Color output
 */
const colors = {
  reset: '\x1b[0m',
  green: '\x1b[32m',
  yellow: '\x1b[33m',
  red: '\x1b[31m',
  cyan: '\x1b[36m'
};

function log(message, color = 'reset') {
  console.error(`${colors[color]}${message}${colors.reset}`);
}

function error(message) {
  log(`  ✗ ${message}`, 'red');
}

function warning(message) {
  log(`  ⚠ ${message}`, 'yellow');
}

function success(message) {
  log(`  ✓ ${message}`, 'green');
}

/**
 * Check 1a: CSP Compliance on NEW content
 * Validates the content being WRITTEN, not the existing file
 */
function checkCSPComplianceOnNewContent() {
  const newContent = HOOK_INPUT.tool_input.content;

  if (!newContent) {
    warning('No content provided in tool input - skipping CSP check');
    return { passed: true };
  }

  // Only check HTML/JS files
  const ext = path.extname(FILE_PATH).toLowerCase();
  if (!['.html', '.js'].includes(ext)) {
    return { passed: true };
  }

  // Exclude scripts/ directory (not served to browsers)
  if (FILE_PATH.includes('/scripts/')) {
    return { passed: true };
  }

  // Exclude email-templates/ directory (email content requires inline styles for client compatibility)
  if (FILE_PATH.includes('/email-templates/')) {
    return { passed: true };
  }

  const violations = [];

  // CSP Violation Patterns
  const patterns = [
    {
      name: 'Inline event handlers',
      regex: /\son\w+\s*=\s*["'][^"']*["']/gi,
      severity: 'CRITICAL'
    },
    {
      name: 'Inline styles',
      regex: /\sstyle\s*=\s*["'][^"']+["']/gi,
      severity: 'CRITICAL'
    },
    {
      name: 'Inline scripts (without src)',
      regex: /<script(?![^>]*\ssrc=)[^>]*>[\s\S]*?<\/script>/gi,
      severity: 'WARNING',
      filter: (match) => match.replace(/<script[^>]*>|<\/script>/gi, '').trim().length > 0
    },
    {
      name: 'javascript: URLs',
      regex: /href\s*=\s*["']javascript:[^"']*["']/gi,
      severity: 'CRITICAL'
    }
  ];

  patterns.forEach(pattern => {
    const matches = newContent.match(pattern.regex);
    if (matches) {
      const filtered = pattern.filter ? matches.filter(pattern.filter) : matches;
      if (filtered.length > 0) {
        violations.push({
          name: pattern.name,
          severity: pattern.severity,
          count: filtered.length,
          samples: filtered.slice(0, 3)
        });
      }
    }
  });

  if (violations.length === 0) {
    return { passed: true };
  }

  // Report violations
  const output = [];
  output.push(`CSP violations detected in NEW content for ${path.basename(FILE_PATH)}:`);
  violations.forEach(v => {
    output.push(`  [${v.severity}] ${v.name} (${v.count} occurrences)`);
    v.samples.forEach((sample, idx) => {
      const truncated = sample.length > 80 ? sample.substring(0, 77) + '...' : sample;
      output.push(`    ${idx + 1}. ${truncated}`);
    });
  });

  return {
    passed: false,
    reason: 'CSP violations in new content',
    output: output.join('\n'),
    violations: violations
  };
}

/**
 * Check 1b: Pre-action-check recency (inst_038)
 */
function checkPreActionCheckRecency() {
  // Only enforce for major file changes
  // Check if pre-action-check was run recently

  try {
    if (!fs.existsSync(SESSION_STATE_PATH)) {
      // No session state, allow but warn
      warning('No session state - pre-action-check recommended (inst_038)');
      return { passed: true };
    }

    const sessionState = JSON.parse(fs.readFileSync(SESSION_STATE_PATH, 'utf8'));
    const lastPreActionCheck = sessionState.last_pre_action_check;
    const currentAction = sessionState.action_count || 0;

    // If never run, just warn (don't block on first action)
    if (!lastPreActionCheck) {
      warning('inst_038: Pre-action-check not run in this session');
      warning('Recommended: node scripts/pre-action-check.js write ' + FILE_PATH);
      return { passed: true }; // Warning only
    }

    // If last run was more than 10 actions ago, warn
    if (currentAction - lastPreActionCheck > 10) {
      warning(`inst_038: Pre-action-check last run ${currentAction - lastPreActionCheck} actions ago`);
      warning('Consider running: node scripts/pre-action-check.js write ' + FILE_PATH);
    }

    return { passed: true };
  } catch (error) {
    warning(`Could not check pre-action-check recency: ${error.message}`);
    return { passed: true };
  }
}

/**
 * Check 2: Overwrite without read
 * CLAUDE.md requires reading files before writing (to avoid overwrites)
 */
function checkOverwriteWithoutRead() {
  if (!fs.existsSync(FILE_PATH)) {
    // New file - no risk of overwrite
    return { passed: true };
  }

  // File exists - this is an overwrite
  // In a real implementation, we'd check if the file was recently read
  // For now, we'll issue a warning but not block
  warning(`File exists - overwriting: ${FILE_PATH}`);
  warning(`Best practice: Read file before writing to avoid data loss`);

  return { passed: true };  // Warning only, not blocking
}

/**
 * Check 3: CrossReferenceValidator
 */
function checkInstructionConflicts() {
  try {
    const content = HOOK_INPUT.tool_input?.content || '';

    const validator = new CrossReferenceValidator({ silent: false });
    const result = validator.validate('file-write', {
      filePath: FILE_PATH,
      content
    });

    if (!result.passed) {
      return {
        passed: false,
        reason: result.conflicts.map(c => c.issue).join('; '),
        conflicts: result.conflicts
      };
    }

    return { passed: true, conflicts: [] };
  } catch (err) {
    warning(`CrossReferenceValidator error: ${err.message}`);
    return { passed: true, conflicts: [] };
  }
}

/**
 * Check 4: BoundaryEnforcer - Values decisions
 */
function checkBoundaryViolation() {
  const valuesIndicators = [
    '/docs/values/',
    '/docs/ethics/',
    'privacy-policy',
    'code-of-conduct',
    'values.html',
    '/pluralistic-values'
  ];

  const isValuesContent = valuesIndicators.some(indicator =>
    FILE_PATH.toLowerCase().includes(indicator.toLowerCase())
  );

  if (isValuesContent) {
    return {
      passed: false,
      reason: 'File appears to contain values content - requires human approval',
      requiresHumanApproval: true
    };
  }

  return { passed: true };
}

/**
 * Update session state
 */
function updateSessionState() {
  try {
    if (fs.existsSync(SESSION_STATE_PATH)) {
      const sessionState = JSON.parse(fs.readFileSync(SESSION_STATE_PATH, 'utf8'));
      sessionState.last_framework_activity = sessionState.last_framework_activity || {};
      sessionState.last_framework_activity.FileWriteHook = {
        timestamp: new Date().toISOString(),
        file: FILE_PATH,
        result: 'passed'
      };
      sessionState.last_updated = new Date().toISOString();
      fs.writeFileSync(SESSION_STATE_PATH, JSON.stringify(sessionState, null, 2));
    }
  } catch (err) {
    // Non-critical
  }
}

/**
 * Main validation
 */
/**
 * Read hook input from stdin
 */
function readHookInput() {
  return new Promise((resolve) => {
    let data = "";
    process.stdin.on("data", chunk => {
      data += chunk;
    });
    process.stdin.on("end", () => {
      try {
        resolve(JSON.parse(data));
      } catch (err) {
        resolve(null);
      }
    });
  });
}

async function main() {
  // Read input from stdin
  HOOK_INPUT = await readHookInput();

  if (!HOOK_INPUT || !HOOK_INPUT.tool_input || !HOOK_INPUT.tool_input.file_path) {
    logMetrics('error', 'No file path in input');
    error('No file path provided in hook input');
    process.exit(2);  // Exit code 2 = BLOCK
  }

  FILE_PATH = HOOK_INPUT.tool_input.file_path;
  log(`\n🔍 Hook: Validating file write: ${FILE_PATH}`, 'cyan');

  // Check 1a: CSP Compliance on NEW content
  const cspCheck = checkCSPComplianceOnNewContent();
  if (!cspCheck.passed) {
    error(cspCheck.reason);
    if (cspCheck.output) {
      console.log(cspCheck.output);
    }
    // Log to audit database
    const cspViolations = (cspCheck.violations || []).map(v => ({
      ruleId: 'inst_038',
      ruleText: v.name,
      severity: v.severity,
      details: `${v.count} occurrences: ${v.samples.slice(0, 2).join('; ')}`
    }));
    await logToAuditDatabase('blocked', cspCheck.reason, cspViolations);
    logMetrics('blocked', cspCheck.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }
  success('CSP compliance validated on new content');

  // Check 1b: Pre-action-check recency (inst_038)
  const preActionCheck = checkPreActionCheckRecency();
  if (!preActionCheck.passed) {
    error(preActionCheck.reason);
    if (preActionCheck.output) {
      console.log(preActionCheck.output);
    }
    await logToAuditDatabase('blocked', preActionCheck.reason, [{
      ruleId: 'inst_038',
      ruleText: 'Pre-action check required before major file modifications',
      severity: 'MEDIUM',
      details: preActionCheck.reason
    }]);
    logMetrics('blocked', preActionCheck.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }
  success('Pre-action-check recency (inst_038) passed');

  // Check 2: Overwrite without read
  const overwriteCheck = checkOverwriteWithoutRead();
  if (!overwriteCheck.passed) {
    error(overwriteCheck.reason);
    await logToAuditDatabase('blocked', overwriteCheck.reason, [{
      ruleId: 'inst_038',
      ruleText: 'Read file before writing to avoid data loss',
      severity: 'MEDIUM',
      details: overwriteCheck.reason
    }]);
    logMetrics('blocked', overwriteCheck.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }

  // Check 3: CrossReferenceValidator
  const conflicts = checkInstructionConflicts();
  if (!conflicts.passed) {
    error(conflicts.reason);
    conflicts.conflicts.forEach(c => {
      log(`    • ${c.id}: ${c.instruction} [${c.quadrant}]`, 'yellow');
    });
    const conflictViolations = conflicts.conflicts.map(c => ({
      ruleId: c.id,
      ruleText: c.instruction,
      severity: 'HIGH',
      details: c.issue
    }));
    await logToAuditDatabase('blocked', conflicts.reason, conflictViolations);
    logMetrics('blocked', conflicts.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }
  success('No instruction conflicts detected');

  // Check 4: BoundaryEnforcer
  const boundary = checkBoundaryViolation();
  if (!boundary.passed) {
    error(boundary.reason);
    await logToAuditDatabase('blocked', boundary.reason, [{
      ruleId: 'inst_020',
      ruleText: 'Values content requires human approval',
      severity: 'CRITICAL',
      details: boundary.reason
    }]);
    logMetrics('blocked', boundary.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }
  success('No boundary violations detected');
  // Check 5: GitHub URL Protection (inst_084)
  const githubUrlCheck = checkGitHubURLProtection();
  if (!githubUrlCheck.passed) {
    error(githubUrlCheck.reason);
    if (githubUrlCheck.output) {
      console.error(githubUrlCheck.output);
    }
    await logToAuditDatabase('blocked', githubUrlCheck.reason, [{
      ruleId: 'inst_084',
      ruleText: 'GitHub URL modifications require explicit approval',
      severity: 'CRITICAL',
      details: githubUrlCheck.reason
    }]);
    logMetrics('blocked', githubUrlCheck.reason);
    process.exit(2);  // Exit code 2 = BLOCK
  }
  success('No unauthorized GitHub URL modifications detected (inst_084)');


  // Update session state
  updateSessionState();

  // Log successful execution to audit database
  await logToAuditDatabase('passed', null, []);
  logMetrics('passed');

  success('File write validation complete\n');
  process.exit(0);
}

main().catch(err => {
  error(`Hook validation error: ${err.message}`);
  logMetrics('error', err.message);
  process.exit(2);  // Exit code 2 = BLOCK
});

/**
 * Log to audit database (MongoDB)
 */
async function logToAuditDatabase(result, reason = null, violations = []) {
  try {
    const mongoose = require('mongoose');

    // Connect to MongoDB
    await mongoose.connect('mongodb://localhost:27017/tractatus_dev', {
      serverSelectionTimeoutMS: 2000
    });

    const AuditLog = require('../../src/models/AuditLog.model');

    const auditEntry = new AuditLog({
      sessionId: HOOK_INPUT.session_id || 'unknown',
      action: 'file_write_validation',
      allowed: result !== 'blocked',
      rulesChecked: violations.map(v => v.ruleId || 'inst_038').filter(Boolean),
      violations: violations.map(v => ({
        ruleId: v.ruleId || 'inst_038',
        ruleText: v.reason || v.name,
        severity: v.severity || 'HIGH',
        details: v.details || v.reason
      })),
      metadata: {
        tool: 'Write',
        file_path: FILE_PATH,
        validation_result: result,
        denial_reason: reason,
        hook: 'validate-file-write'
      },
      domain: 'SYSTEM',
      service: 'FileWriteValidator',
      environment: 'development',
      timestamp: new Date()
    });

    await auditEntry.save();
    await mongoose.disconnect();
  } catch (err) {
    // Non-fatal: Continue even if logging fails
    console.error('[Audit Logging] Failed:', err.message);
  }
}

/**
 * Log metrics for hook execution
 */
function logMetrics(result, reason = null) {
  try {
    const METRICS_PATH = path.join(__dirname, '../../.claude/metrics/hooks-metrics.json');
    const METRICS_DIR = path.dirname(METRICS_PATH);
    if (!fs.existsSync(METRICS_DIR)) {
      fs.mkdirSync(METRICS_DIR, { recursive: true });
    }
    let metrics = { hook_executions: [], blocks: [], session_stats: {} };
    if (fs.existsSync(METRICS_PATH)) {
      metrics = JSON.parse(fs.readFileSync(METRICS_PATH, 'utf8'));
    }
    metrics.hook_executions.push({
      hook: 'validate-file-write',
      timestamp: new Date().toISOString(),
      file: FILE_PATH,
      result: result,
      reason: reason
    });
    if (result === 'blocked') {
      metrics.blocks.push({
        hook: 'validate-file-write',
        timestamp: new Date().toISOString(),
        file: FILE_PATH,
        reason: reason
      });
    }
    metrics.session_stats.total_write_hooks = (metrics.session_stats.total_write_hooks || 0) + 1;
    metrics.session_stats.total_write_blocks = (metrics.session_stats.total_write_blocks || 0) + (result === 'blocked' ? 1 : 0);
    metrics.session_stats.last_updated = new Date().toISOString();
    if (metrics.hook_executions.length > 1000) {
      metrics.hook_executions = metrics.hook_executions.slice(-1000);
    }
    if (metrics.blocks.length > 500) {
      metrics.blocks = metrics.blocks.slice(-500);
    }
    fs.writeFileSync(METRICS_PATH, JSON.stringify(metrics, null, 2));
  } catch (err) {
    // Non-critical
  }
}

/**
 * Check 5: GitHub URL Protection (inst_084) - HARD BLOCK unauthorized changes
 */
function checkGitHubURLProtection() {
  const oldString = HOOK_INPUT.tool_input?.old_string || '';
  const newString = HOOK_INPUT.tool_input?.new_string || '';

  // Only check if both strings contain github.com
  if (!oldString.includes('github.com') && !newString.includes('github.com')) {
    return { passed: true };
  }

  // Extract GitHub repository URLs (github.com/org/repo format)
  const githubUrlPattern = /github\.com\/[\w-]+\/[\w-]+/g;
  const oldUrls = oldString.match(githubUrlPattern) || [];
  const newUrls = newString.match(githubUrlPattern) || [];

  // Check 1: New GitHub URLs added (requires approval)
  if (newUrls.length > oldUrls.length) {
    return {
      passed: false,
      reason: 'inst_084: New GitHub URL detected - requires explicit approval',
      output: `Attempting to add GitHub URL:\n  New: ${newUrls.join(', ')}\n  Old: ${oldUrls.join(', ')}\n\nBLOCKED: Adding GitHub URLs requires explicit human approval to prevent exposure of private repository structure.`
    };
  }

  // Check 2: GitHub repository name changed (HIGH RISK)
  if (oldUrls.length > 0 && newUrls.length > 0) {
    const oldRepos = new Set(oldUrls);
    const newRepos = new Set(newUrls);

    // Find any repository name changes
    for (const oldRepo of oldRepos) {
      if (!newRepos.has(oldRepo)) {
        return {
          passed: false,
          reason: 'inst_084: GitHub repository name change detected - HARD BLOCK',
          output: `CRITICAL: Repository URL modification detected:\n  Old: ${oldRepo}\n  New: ${Array.from(newRepos).join(', ')}\n\nBLOCKED: This could expose private repository structure (e.g., tractatus-framework → tractatus exposes /src/services/).\n\nRequired: Explicit human approval for ANY GitHub repository URL changes.`
        };
      }
    }
  }

  // Check 3: GitHub URL path changed (e.g., /tree/main/src/services → /tree/main/deployment-quickstart)
  const oldPaths = oldString.match(/github\.com[^\s"'<>]*/g) || [];
  const newPaths = newString.match(/github\.com[^\s"'<>]*/g) || [];

  if (oldPaths.length > 0 && newPaths.length > 0) {
    for (let i = 0; i < Math.min(oldPaths.length, newPaths.length); i++) {
      if (oldPaths[i] !== newPaths[i]) {
        return {
          passed: false,
          reason: 'inst_084: GitHub URL path modification detected - requires approval',
          output: `GitHub URL path change detected:\n  Old: ${oldPaths[i]}\n  New: ${newPaths[i]}\n\nBLOCKED: URL path modifications require explicit approval to prevent linking to non-existent or private paths.`
        };
      }
    }
  }

  return { passed: true };
}