@ryuenn3123/agentic-senior-core 4.4.0 → 5.2.0

package/scripts/validate/audits/rule-id-uniqueness.mjs

@@ -1,278 +0,0 @@
-#!/usr/bin/env node
-// @ts-check
-
-/**
- * audit-rule-id-uniqueness.mjs
- *
- * Phase 1 GATE B citability gate. Checks every migrated rule file under
- * `.agent-context/rules/` for:
- *
- *   1. YAML frontmatter parses and contains required keys (id_prefix, domain,
- *      priority, scope, applies_to, keywords).
- *   2. Section IDs match the file's locked id_prefix and are unique within
- *      the file.
- *   3. Every `[REF:<PREFIX>-NNN]` mention across the rules pack, prompts/,
- *      and review-checklists/ resolves to a real section ID.
- *   4. Every `related:` entry in any rule's frontmatter resolves to a real
- *      section ID.
- *   5. No ambiguous prose references (`see above`, `as noted earlier`,
- *      `the next section`, etc.) survive in any rule body.
- *
- * Files that have not been migrated yet (no YAML frontmatter) are skipped
- * with a notice. The audit only enforces shape on migrated files.
- *
- * Usage:
- *   node scripts/audit-rule-id-uniqueness.mjs            (human + report line)
- *   node scripts/audit-rule-id-uniqueness.mjs --json     (JSON only)
- *
- * Exit codes:
- *   0 — clean
- *   1 — at least one violation
- */
-
-import { readdirSync, readFileSync, existsSync } from 'node:fs';
-import { dirname, join, resolve } from 'node:path';
-import { fileURLToPath } from 'node:url';
-import { parse as parseYaml } from 'yaml';
-
-const SCRIPT_FILE_PATH = fileURLToPath(import.meta.url);
-const REPOSITORY_ROOT = resolve(dirname(SCRIPT_FILE_PATH), '../../..');
-const RULES_DIRECTORY = join(REPOSITORY_ROOT, '.agent-context', 'rules');
-const PROMPTS_DIRECTORY = join(REPOSITORY_ROOT, '.agent-context', 'prompts');
-const REVIEW_CHECKLISTS_DIRECTORY = join(REPOSITORY_ROOT, '.agent-context', 'review-checklists');
-
-const ARGS = new Set(process.argv.slice(2));
-const JSON_ONLY = ARGS.has('--json');
-
-const RULE_ID_PATTERN = /\b([A-Z]+)-(\d{3,4})(?:-([A-Z]))?\b/;
-const REF_PATTERN = /\[REF:([A-Z]+-\d{3,4}(?:-[A-Z])?)\]/g;
-const SECTION_HEADING_PATTERN = /^##\s+([A-Z]+-\d{3,4}(?:-[A-Z])?):\s+(.+)$/;
-const REQUIRED_FRONTMATTER_KEYS = ['id_prefix', 'domain', 'priority', 'scope', 'applies_to', 'keywords'];
-const FRONTMATTER_PATTERN = /^---\n([\s\S]*?)\n---\n/;
-
-const AMBIGUOUS_PROSE_REFERENCE_PATTERNS = [
-  /\bsee above\b/i,
-  /\bsee below\b/i,
-  /\bas noted earlier\b/i,
-  /\bas noted above\b/i,
-  /\bas mentioned earlier\b/i,
-  /\bas mentioned above\b/i,
-  /\bin the previous section\b/i,
-  /\bin the next section\b/i,
-  /\bthe next section\b/i,
-  /\bthe previous section\b/i,
-  /\bsee earlier\b/i,
-  /\bsee later\b/i,
-];
-
-function listRuleFiles() {
-  return readdirSync(RULES_DIRECTORY)
-    .filter((filename) => filename.endsWith('.md') && !filename.endsWith('.candidate.md'))
-    .sort()
-    .map((filename) => ({ filename, absolutePath: join(RULES_DIRECTORY, filename) }));
-}
-
-function parseFrontmatter(sourceText) {
-  const match = sourceText.match(FRONTMATTER_PATTERN);
-  if (!match) return null;
-  try {
-    return parseYaml(match[1]);
-  } catch {
-    return null;
-  }
-}
-
-function extractSectionIds(sourceText) {
-  const sectionIds = [];
-  for (const line of sourceText.split(/\r?\n/)) {
-    const match = line.match(SECTION_HEADING_PATTERN);
-    if (match) {
-      sectionIds.push({ id: match[1], title: match[2].trim() });
-    }
-  }
-  return sectionIds;
-}
-
-function findAmbiguousProseReferences(sourceText) {
-  const findings = [];
-  const lines = sourceText.split(/\r?\n/);
-  for (let lineIndex = 0; lineIndex < lines.length; lineIndex += 1) {
-    for (const pattern of AMBIGUOUS_PROSE_REFERENCE_PATTERNS) {
-      if (pattern.test(lines[lineIndex])) {
-        findings.push({ lineNumber: lineIndex + 1, snippet: lines[lineIndex].trim() });
-        break;
-      }
-    }
-  }
-  return findings;
-}
-
-function listMarkdownFilesIn(absoluteDirectoryPath) {
-  if (!existsSync(absoluteDirectoryPath)) return [];
-  return readdirSync(absoluteDirectoryPath)
-    .filter((filename) => filename.endsWith('.md'))
-    .map((filename) => join(absoluteDirectoryPath, filename));
-}
-
-function collectAllRefMentions() {
-  const mentions = [];
-  const candidatePaths = [
-    ...listRuleFiles().map((entry) => entry.absolutePath),
-    ...listMarkdownFilesIn(PROMPTS_DIRECTORY),
-    ...listMarkdownFilesIn(REVIEW_CHECKLISTS_DIRECTORY),
-  ];
-  for (const filePath of candidatePaths) {
-    const sourceText = readFileSync(filePath, 'utf8');
-    for (const match of sourceText.matchAll(REF_PATTERN)) {
-      mentions.push({ filePath, refId: match[1] });
-    }
-  }
-  return mentions;
-}
-
-export function runRuleIdUniquenessAudit() {
-  const violations = [];
-  const ruleFiles = listRuleFiles();
-  const allKnownSectionIds = new Set();
-  const migratedFileCount = { migrated: 0, skipped: 0 };
-  const perFile = [];
-
-  for (const { filename, absolutePath } of ruleFiles) {
-    const sourceText = readFileSync(absolutePath, 'utf8');
-    const frontmatter = parseFrontmatter(sourceText);
-    if (!frontmatter) {
-      migratedFileCount.skipped += 1;
-      perFile.push({ filename, status: 'skipped', reason: 'no-yaml-frontmatter' });
-      continue;
-    }
-
-    migratedFileCount.migrated += 1;
-
-    for (const requiredKey of REQUIRED_FRONTMATTER_KEYS) {
-      if (!(requiredKey in frontmatter)) {
-        violations.push({
-          file: filename,
-          kind: 'frontmatter.missing-key',
-          detail: `frontmatter is missing required key: ${requiredKey}`,
-        });
-      }
-    }
-
-    const idPrefix = frontmatter.id_prefix;
-    if (typeof idPrefix !== 'string' || idPrefix.length === 0) {
-      violations.push({ file: filename, kind: 'frontmatter.invalid-id-prefix', detail: 'id_prefix must be a non-empty string' });
-      perFile.push({ filename, status: 'failed', reason: 'invalid-id-prefix' });
-      continue;
-    }
-
-    const sections = extractSectionIds(sourceText);
-    const seenIdsInFile = new Set();
-    for (const section of sections) {
-      const idMatch = section.id.match(RULE_ID_PATTERN);
-      if (!idMatch || idMatch[1] !== idPrefix) {
-        violations.push({
-          file: filename,
-          kind: 'id.prefix-mismatch',
-          detail: `section id '${section.id}' does not match the file's id_prefix '${idPrefix}'`,
-        });
-        continue;
-      }
-      if (seenIdsInFile.has(section.id)) {
-        violations.push({
-          file: filename,
-          kind: 'id.duplicate',
-          detail: `section id '${section.id}' is reused within the same file`,
-        });
-      }
-      seenIdsInFile.add(section.id);
-      allKnownSectionIds.add(section.id);
-    }
-
-    const ambiguousProse = findAmbiguousProseReferences(sourceText);
-    for (const finding of ambiguousProse) {
-      violations.push({
-        file: filename,
-        kind: 'prose.ambiguous-reference',
-        detail: `line ${finding.lineNumber}: ${finding.snippet}`,
-      });
-    }
-
-    perFile.push({
-      filename,
-      status: 'audited',
-      idPrefix,
-      sectionCount: sections.length,
-      ambiguousProseCount: ambiguousProse.length,
-      knownSectionIdsInFile: sections.map((section) => section.id),
-      relatedRefs: (frontmatter.related && typeof frontmatter.related === 'object' && !Array.isArray(frontmatter.related))
-        ? frontmatter.related
-        : null,
-    });
-  }
-
-  for (const fileEntry of perFile) {
-    if (fileEntry.status !== 'audited') continue;
-    if (!fileEntry.relatedRefs) continue;
-    for (const [parentId, relatedList] of Object.entries(fileEntry.relatedRefs)) {
-      if (!fileEntry.knownSectionIdsInFile.includes(parentId)) {
-        violations.push({
-          file: fileEntry.filename,
-          kind: 'related.parent-missing',
-          detail: `related map key '${parentId}' is not a section in this file`,
-        });
-      }
-      if (!Array.isArray(relatedList)) {
-        violations.push({
-          file: fileEntry.filename,
-          kind: 'related.malformed',
-          detail: `related[${parentId}] must be an array of <PREFIX>-NNN ids`,
-        });
-        continue;
-      }
-      for (const relatedId of relatedList) {
-        if (typeof relatedId !== 'string' || !RULE_ID_PATTERN.test(relatedId)) {
-          violations.push({
-            file: fileEntry.filename,
-            kind: 'related.malformed',
-            detail: `related[${parentId}] entry '${relatedId}' is not a valid <PREFIX>-NNN id`,
-          });
-          continue;
-        }
-        if (!allKnownSectionIds.has(relatedId)) {
-          violations.push({
-            file: fileEntry.filename,
-            kind: 'related.unresolved',
-            detail: `related[${parentId}] entry '${relatedId}' does not resolve to any section in the rules pack`,
-          });
-        }
-      }
-    }
-  }
-
-  const refMentions = collectAllRefMentions();
-  for (const mention of refMentions) {
-    if (!allKnownSectionIds.has(mention.refId)) {
-      violations.push({
-        file: mention.filePath.replace(REPOSITORY_ROOT, '').replace(/^[\\/]+/, ''),
-        kind: 'ref.unresolved',
-        detail: `[REF:${mention.refId}] does not resolve to any section in the rules pack`,
-      });
-    }
-  }
-
-  return {
-    auditName: 'audit-rule-id-uniqueness',
-    reportVersion: '1.0.0',
-    generatedAt: new Date().toISOString(),
-    migratedFileCount: migratedFileCount.migrated,
-    skippedFileCount: migratedFileCount.skipped,
-    knownSectionIdCount: allKnownSectionIds.size,
-    refMentionCount: refMentions.length,
-    perFile,
-    violationCount: violations.length,
-    violations,
-    passed: violations.length === 0,
-  };
-}
-
-