@howlil/ez-agents 3.4.2 → 3.5.0

package/ez-agents/bin/lib/discussion-synthesizer.cjs

@@ -0,0 +1,458 @@
+#!/usr/bin/env node
+
+/**
+ * Discussion Synthesizer — Reads DISCUSSION.md, extracts consensus and blockers
+ *
+ * Parses the multi-agent DISCUSSION.md format to extract:
+ * - Hard blockers from any agent
+ * - Warnings and advisory notes
+ * - Consensus status (open | consensus-reached | needs-human)
+ * - Go/No-Go recommendation
+ */
+
+'use strict';
+
+const fs = require('fs');
+const path = require('path');
+const { withLock } = require('./file-lock.cjs');
+
+// ─────────────────────────────────────────────
+// Parser
+// ─────────────────────────────────────────────
+
+/**
+ * Parse DISCUSSION.md file
+ * @param {string} filePath
+ * @returns {{ frontmatter: object, sections: object[], consensus: object, blockers: string[], warnings: string[] }}
+ */
+function parseDiscussion(filePath) {
+  if (!fs.existsSync(filePath)) {
+    return {
+      found: false,
+      filePath,
+      frontmatter: {},
+      sections: [],
+      consensus: { status: 'open', goNoGo: 'GO', rationale: 'No discussion file — proceeding' },
+      blockers: [],
+      warnings: []
+    };
+  }
+
+  const content = fs.readFileSync(filePath, 'utf8');
+  const lines = content.split('\n');
+
+  // Parse YAML frontmatter
+  const frontmatter = parseFrontmatter(content);
+
+  // Parse agent sections
+  const sections = parseAgentSections(lines);
+
+  // Extract blockers and warnings from all sections
+  const blockers = [];
+  const warnings = [];
+
+  for (const section of sections) {
+    const sectionBlockers = extractBlockers(section.content);
+    const sectionWarnings = extractWarnings(section.content);
+    blockers.push(...sectionBlockers.map(b => ({ agent: section.agent, text: b })));
+    warnings.push(...sectionWarnings.map(w => ({ agent: section.agent, text: w })));
+  }
+
+  // Parse consensus section
+  const consensus = parseConsensus(sections);
+
+  return {
+    found: true,
+    filePath,
+    frontmatter,
+    sections,
+    consensus,
+    blockers,
+    warnings,
+    hasBlockers: blockers.length > 0,
+    hasWarnings: warnings.length > 0
+  };
+}
+
+/**
+ * Parse YAML frontmatter from discussion file
+ */
+function parseFrontmatter(content) {
+  const match = content.match(/^---\n([\s\S]*?)\n---/);
+  if (!match) return {};
+
+  const fm = {};
+  const lines = match[1].split('\n');
+  for (const line of lines) {
+    const colonIdx = line.indexOf(':');
+    if (colonIdx === -1) continue;
+    const key = line.slice(0, colonIdx).trim();
+    const value = line.slice(colonIdx + 1).trim();
+    // Remove quotes
+    fm[key] = value.replace(/^['"]|['"]$/g, '');
+  }
+  return fm;
+}
+
+/**
+ * Parse agent sections from DISCUSSION.md
+ * Returns array of { agent, heading, content }
+ */
+function parseAgentSections(lines) {
+  const sections = [];
+  let currentSection = null;
+  let currentContent = [];
+
+  const agentHeadings = [
+    { pattern: /## Requirements Perspective/, agent: 'requirements' },
+    { pattern: /## Tech Lead Perspective/, agent: 'tech-lead' },
+    { pattern: /## Observer Perspective/, agent: 'observer' },
+    { pattern: /## Scrum Master Perspective/, agent: 'scrum-master' },
+    { pattern: /## Consensus/, agent: 'consensus' }
+  ];
+
+  for (let i = 0; i < lines.length; i++) {
+    const line = lines[i];
+
+    // Check if line starts a known section
+    const matchedHeading = agentHeadings.find(h => h.pattern.test(line));
+
+    if (matchedHeading) {
+      // Save previous section
+      if (currentSection) {
+        sections.push({
+          agent: currentSection.agent,
+          heading: currentSection.heading,
+          content: currentContent.join('\n').trim()
+        });
+      }
+      currentSection = { agent: matchedHeading.agent, heading: line };
+      currentContent = [];
+    } else if (currentSection) {
+      currentContent.push(line);
+    }
+  }
+
+  // Push last section
+  if (currentSection) {
+    sections.push({
+      agent: currentSection.agent,
+      heading: currentSection.heading,
+      content: currentContent.join('\n').trim()
+    });
+  }
+
+  return sections;
+}
+
+/**
+ * Extract blocker statements from section content
+ * Looks for BLOCKER markers used by agents
+ */
+function extractBlockers(content) {
+  const blockers = [];
+  const lines = content.split('\n');
+
+  for (const line of lines) {
+    if (line.includes('🛑') || line.includes('BLOCKER') || line.match(/\*\*BLOCKER/i)) {
+      // Extract the description after the marker
+      const cleaned = line
+        .replace(/[🛑*]/g, '')
+        .replace(/BLOCKER\s*—?\s*/i, '')
+        .trim();
+      if (cleaned && cleaned.length > 3) {
+        blockers.push(cleaned);
+      }
+    }
+  }
+
+  return blockers;
+}
+
+/**
+ * Extract warning statements from section content
+ */
+function extractWarnings(content) {
+  const warnings = [];
+  const lines = content.split('\n');
+
+  for (const line of lines) {
+    if (line.includes('⚠️') || line.match(/\*\*WARNING/i)) {
+      const cleaned = line
+        .replace(/[⚠️*]/g, '')
+        .replace(/WARNING\s*—?\s*/i, '')
+        .trim();
+      if (cleaned && cleaned.length > 3) {
+        warnings.push(cleaned);
+      }
+    }
+  }
+
+  return warnings;
+}
+
+/**
+ * Parse consensus section for Go/No-Go status
+ */
+function parseConsensus(sections) {
+  const consensusSection = sections.find(s => s.agent === 'consensus');
+
+  if (!consensusSection) {
+    return { status: 'open', goNoGo: 'GO', rationale: 'No consensus section yet' };
+  }
+
+  const content = consensusSection.content;
+
+  // Extract Go/No-Go
+  let goNoGo = 'GO';
+  if (content.match(/NO-GO/i)) goNoGo = 'NO-GO';
+  else if (content.match(/HUMAN-NEEDED/i)) goNoGo = 'HUMAN-NEEDED';
+  else if (content.match(/^.*GO.*$/m)) goNoGo = 'GO';
+
+  // Extract status
+  let status = 'open';
+  if (content.match(/consensus-reached/i) || goNoGo !== 'open') status = 'consensus-reached';
+  if (goNoGo === 'HUMAN-NEEDED') status = 'needs-human';
+
+  // Extract rationale
+  const rationaleMatch = content.match(/### Rationale\n([^\n]+)/);
+  const rationale = rationaleMatch ? rationaleMatch[1].trim() : '';
+
+  return { status, goNoGo, rationale };
+}
+
+// ─────────────────────────────────────────────
+// Synthesis
+// ─────────────────────────────────────────────
+
+/**
+ * Synthesize discussion into orchestrator-ready decision
+ * @param {string} discussionPath - Path to DISCUSSION.md
+ * @returns {{ proceed: boolean, reason: string, blockers: object[], warnings: object[], score: object }}
+ */
+function synthesize(discussionPath) {
+  const discussion = parseDiscussion(discussionPath);
+
+  if (!discussion.found) {
+    return {
+      proceed: true,
+      reason: 'No discussion file — no pre-flight concerns',
+      blockers: [],
+      warnings: [],
+      agentsParticipated: []
+    };
+  }
+
+  const hasBlockers = discussion.blockers.length > 0;
+  const consensusGoNoGo = discussion.consensus.goNoGo;
+
+  // Determine whether to proceed
+  let proceed = true;
+  let reason = '';
+
+  if (hasBlockers || consensusGoNoGo === 'NO-GO') {
+    proceed = false;
+    reason = hasBlockers
+      ? `${discussion.blockers.length} blocker(s) must be resolved before execution`
+      : 'Consensus is NO-GO';
+  } else if (consensusGoNoGo === 'HUMAN-NEEDED') {
+    proceed = false;
+    reason = 'Human input required before proceeding';
+  } else {
+    proceed = true;
+    reason = discussion.warnings.length > 0
+      ? `${discussion.warnings.length} warning(s) — proceeding with awareness`
+      : 'No blockers found — proceeding';
+  }
+
+  const agentsParticipated = discussion.sections
+    .filter(s => s.agent !== 'consensus')
+    .filter(s => !s.content.includes('{Populated') && s.content.trim().length > 20)
+    .map(s => s.agent);
+
+  return {
+    proceed,
+    reason,
+    blockers: discussion.blockers,
+    warnings: discussion.warnings,
+    consensus: discussion.consensus,
+    agentsParticipated,
+    frontmatter: discussion.frontmatter
+  };
+}
+
+/**
+ * Check if a discussion file needs updating (agents haven't written yet)
+ * @param {string} discussionPath
+ * @returns {{ needsObserver: boolean, needsTechLead: boolean, needsScrumMaster: boolean }}
+ */
+function checkParticipation(discussionPath) {
+  const discussion = parseDiscussion(discussionPath);
+
+  if (!discussion.found) {
+    return { needsObserver: true, needsTechLead: true, needsScrumMaster: true };
+  }
+
+  const populated = (agent) => {
+    const section = discussion.sections.find(s => s.agent === agent);
+    return section && !section.content.includes('{Populated') && section.content.trim().length > 20;
+  };
+
+  return {
+    needsObserver: !populated('observer'),
+    needsTechLead: !populated('tech-lead'),
+    needsScrumMaster: !populated('scrum-master'),
+    needsRequirements: !populated('requirements')
+  };
+}
+
+/**
+ * Update consensus section in DISCUSSION.md
+ * @param {string} discussionPath
+ * @param {object} consensusData - { goNoGo, blockers, warnings, rationale }
+ */
+async function updateConsensus(discussionPath, consensusData) {
+  if (!fs.existsSync(discussionPath)) return false;
+  return withLock(discussionPath, async () => {
+    const content = fs.readFileSync(discussionPath, 'utf8');
+
+    const blockerList = consensusData.blockers.length > 0
+      ? consensusData.blockers.map(b => `- 🛑 ${b.agent}: ${b.text}`).join('\n')
+      : 'None';
+
+    const warningList = consensusData.warnings.length > 0
+      ? consensusData.warnings.slice(0, 5).map(w => `- ⚠️ ${w.agent}: ${w.text}`).join('\n')
+      : 'None';
+
+    const now = new Date().toISOString();
+    const status = consensusData.goNoGo === 'GO'
+      ? 'consensus-reached'
+      : consensusData.goNoGo === 'HUMAN-NEEDED' ? 'needs-human' : 'consensus-reached';
+
+    const consensusSection = `## Consensus
+
+> *Synthesized by orchestrator from above perspectives*
+
+**Status:** ${status}
+
+### Blockers
+${blockerList}
+
+### Key Warnings
+${warningList}
+
+### Go / No-Go
+${consensusData.goNoGo} — ${consensusData.rationale}
+
+### Rationale
+${consensusData.rationale}
+
+---
+
+*Discussion opened: {timestamp}*
+*Last updated: ${now}*`;
+
+    // Replace existing consensus section or append
+    let updated;
+    if (content.includes('## Consensus')) {
+      updated = content.replace(/## Consensus[\s\S]*$/, consensusSection);
+    } else {
+      updated = content + '\n\n' + consensusSection;
+    }
+
+    // Also update frontmatter status
+    updated = updated.replace(/^status: .*$/m, `status: ${status}`);
+
+    fs.writeFileSync(discussionPath, updated, 'utf8');
+    return true;
+  });
+}
+
+/**
+ * Format synthesis result as human-readable text
+ * @param {object} result - From synthesize()
+ * @returns {string}
+ */
+function formatSynthesis(result) {
+  const lines = [];
+
+  lines.push(`## Pre-Flight Discussion Summary`);
+  lines.push(`**Decision:** ${result.proceed ? '✓ GO — proceed to execution' : '✗ ' + (result.consensus && result.consensus.goNoGo === 'HUMAN-NEEDED' ? 'HUMAN-NEEDED' : 'NO-GO')}`);
+  lines.push(`**Reason:** ${result.reason}`);
+  lines.push(`**Agents participated:** ${result.agentsParticipated.join(', ') || 'none'}`);
+
+  if (result.blockers.length > 0) {
+    lines.push('');
+    lines.push('### Blockers (must resolve)');
+    for (const b of result.blockers) {
+      lines.push(`- 🛑 **${b.agent}:** ${b.text}`);
+    }
+  }
+
+  if (result.warnings.length > 0) {
+    lines.push('');
+    lines.push('### Warnings (advisory)');
+    for (const w of result.warnings.slice(0, 5)) {
+      lines.push(`- ⚠️ **${w.agent}:** ${w.text}`);
+    }
+  }
+
+  return lines.join('\n');
+}
+
+// ─────────────────────────────────────────────
+// CLI Interface
+// ─────────────────────────────────────────────
+
+if (require.main === module) {
+  const args = process.argv.slice(2);
+  const cmd = args[0];
+  const discussionPath = args[1];
+
+  if (!cmd) {
+    console.error('Usage: discussion-synthesizer.cjs <synthesize|check-participation|update-consensus> <path> [options]');
+    process.exit(1);
+  }
+
+  (async () => {
+  try {
+    if (cmd === 'synthesize') {
+      if (!discussionPath) { console.error('Path required'); process.exit(1); }
+      const result = synthesize(discussionPath);
+      if (args.includes('--json')) {
+        console.log(JSON.stringify(result, null, 2));
+      } else {
+        console.log(formatSynthesis(result));
+        process.exit(result.proceed ? 0 : 1);
+      }
+    } else if (cmd === 'check-participation') {
+      if (!discussionPath) { console.error('Path required'); process.exit(1); }
+      const result = checkParticipation(discussionPath);
+      console.log(JSON.stringify(result, null, 2));
+    } else if (cmd === 'update-consensus') {
+      if (!discussionPath) { console.error('Path required'); process.exit(1); }
+      const dataArg = args[2];
+      if (!dataArg) { console.error('Consensus data JSON required'); process.exit(1); }
+      const data = JSON.parse(dataArg);
+      const ok = await updateConsensus(discussionPath, data);
+      console.log(JSON.stringify({ updated: ok }));
+    } else {
+      console.error(`Unknown command: ${cmd}`);
+      process.exit(1);
+    }
+  } catch (err) {
+    console.error(`Error: ${err.message}`);
+    process.exit(1);
+  }
+  })();
+}
+
+module.exports = {
+  parseDiscussion,
+  synthesize,
+  checkParticipation,
+  updateConsensus,
+  formatSynthesis,
+  extractBlockers,
+  extractWarnings
+};

package/ez-agents/bin/lib/file-access.cjs

@@ -0,0 +1,207 @@
+#!/usr/bin/env node
+
+/**
+ * File Access Service
+ *
+ * Provides file reading capabilities with glob pattern support.
+ * Uses micromatch for glob matching with support for negation and brace expansion.
+ */
+
+const fs = require('fs');
+const path = require('path');
+const micromatch = require('micromatch');
+const { FileAccessError } = require('./context-errors.cjs');
+
+const MAX_FILE_COUNT = 1000;
+
+class FileAccessService {
+  /**
+   * Create a new FileAccessService instance
+   * @param {string} cwd - Current working directory (defaults to process.cwd())
+   */
+  constructor(cwd) {
+    this.cwd = cwd || process.cwd();
+  }
+
+  /**
+   * Read files matching patterns
+   * @param {string|string[]} patterns - File patterns (glob or single path)
+   * @returns {Array<{path: string, content: string}>} - Array of file objects
+   * @throws {FileAccessError} - On file access errors
+   */
+  readFiles(patterns) {
+    const patternArray = Array.isArray(patterns) ? patterns : [patterns];
+    
+    // Get all files recursively
+    const allFiles = this._getAllFiles(this.cwd);
+    
+    // Convert paths to relative paths from cwd
+    const relativeFiles = allFiles.map(f => {
+      const relPath = path.relative(this.cwd, f);
+      // Convert to POSIX style paths for glob matching
+      return relPath.replace(/\\/g, '/');
+    });
+    
+    // Filter with micromatch
+    const matchedFiles = micromatch.match(relativeFiles, patternArray, {
+      dot: false, // Don't match hidden files/directories by default
+      nocase: false
+    });
+    
+    // Check max file count
+    if (matchedFiles.length > MAX_FILE_COUNT) {
+      throw new FileAccessError(
+        matchedFiles[0],
+        `Max file count exceeded: ${matchedFiles.length} > ${MAX_FILE_COUNT}`
+      );
+    }
+    
+    // Read file contents
+    const results = matchedFiles.map(filePath => {
+      const fullPath = path.join(this.cwd, filePath);
+      
+      if (!fs.existsSync(fullPath)) {
+        throw new FileAccessError(filePath, 'File not found');
+      }
+      
+      try {
+        const content = fs.readFileSync(fullPath, 'utf-8');
+        return {
+          path: filePath,
+          content
+        };
+      } catch (err) {
+        if (err.code === 'EACCES') {
+          throw new FileAccessError(filePath, 'Permission denied');
+        }
+        throw new FileAccessError(filePath, err.message);
+      }
+    });
+    
+    return results;
+  }
+
+  /**
+   * Get all files recursively from a directory
+   * @param {string} dir - Directory to scan
+   * @returns {string[]} - Array of file paths
+   * @private
+   */
+  _getAllFiles(dir) {
+    const files = [];
+    
+    try {
+      const entries = fs.readdirSync(dir, { withFileTypes: true });
+      
+      for (const entry of entries) {
+        const fullPath = path.join(dir, entry.name);
+        
+        // Skip hidden directories (starting with .)
+        if (entry.isDirectory() && !entry.name.startsWith('.')) {
+          const subFiles = this._getAllFiles(fullPath);
+          files.push(...subFiles);
+        } else if (entry.isFile()) {
+          files.push(fullPath);
+        }
+      }
+    } catch (err) {
+      // Ignore permission errors during directory traversal
+      if (err.code !== 'EACCES') {
+        throw err;
+      }
+    }
+    
+    return files;
+  }
+
+  /**
+   * Normalize a file path (convert Windows paths to Unix style)
+   * @param {string} filePath - The path to normalize
+   * @returns {string} - Normalized path
+   * @throws {FileAccessError} - On path traversal attempts
+   */
+  normalizePath(filePath) {
+    // Convert Windows backslashes to forward slashes
+    const normalized = filePath.replace(/\\/g, '/');
+    
+    // Check for path traversal attempts
+    if (normalized.includes('..')) {
+      throw new FileAccessError(filePath, 'Path traversal not allowed');
+    }
+    
+    return normalized;
+  }
+
+  /**
+   * Validate a file path
+   * @param {string} filePath - The path to validate
+   * @returns {boolean} - True if valid
+   */
+  validatePath(filePath) {
+    // Reject paths with null bytes
+    if (filePath.includes('\x00')) {
+      return false;
+    }
+    
+    // Resolve the path
+    const resolvedPath = path.resolve(this.cwd, filePath);
+    
+    // Check if path is within cwd (prevent access outside project)
+    const normalizedCwd = path.resolve(this.cwd).replace(/\\/g, '/');
+    const normalizedResolved = resolvedPath.replace(/\\/g, '/');
+    
+    // Path must be within or equal to cwd
+    if (!normalizedResolved.startsWith(normalizedCwd)) {
+      return false;
+    }
+    
+    return true;
+  }
+
+  /**
+   * Check if a file exists
+   * @param {string} filePath - Path to check
+   * @returns {boolean} - True if file exists
+   */
+  fileExists(filePath) {
+    const fullPath = path.join(this.cwd, filePath);
+    return fs.existsSync(fullPath);
+  }
+
+  /**
+   * Read a single file
+   * @param {string} filePath - Path to the file
+   * @returns {{path: string, content: string}} - File object
+   * @throws {FileAccessError} - On file access errors
+   */
+  readFile(filePath) {
+    const results = this.readFiles([filePath]);
+    return results[0] || null;
+  }
+
+  /**
+   * Get file info (size, modified time, etc.)
+   * @param {string} filePath - Path to the file
+   * @returns {{path: string, size: number, mtime: Date}} - File info
+   * @throws {FileAccessError} - On file access errors
+   */
+  getFileInfo(filePath) {
+    const fullPath = path.join(this.cwd, filePath);
+    
+    if (!fs.existsSync(fullPath)) {
+      throw new FileAccessError(filePath, 'File not found');
+    }
+    
+    const stats = fs.statSync(fullPath);
+    
+    return {
+      path: filePath,
+      size: stats.size,
+      mtime: stats.mtime,
+      isDirectory: stats.isDirectory(),
+      isFile: stats.isFile()
+    };
+  }
+}
+
+module.exports = FileAccessService;