okr/scripts/okf-check.mjs

#!/usr/bin/env node
// okf-check.mjs
// Validerer en OKF-bundle-rot. Kontrakt (de-facto OKF: kun `type` paakrevd):
//   - Hver konsept-fil (.md unntatt index.md) MAA ha `type:` i frontmatter.
//   - >= 1 fil uten type  -> exit 1 + teller + navngir filene.
//   - 0 filer uten type   -> exit 0.
// Anbefalte felt (resource/title/description/timestamp) rapporteres som ADVARSEL,
// ikke feil. Rotens `okf_version` ekkoes for menneskelig sammenligning mot
// gjeldende standard (ingen auto-fetch — hooks/scripts er no-network; SC7 myket).
//
// Kjoeres PER ROT (prosjekt `.claude/okr/` + home `~/.claude/okr/org/`).
// Zero npm dependencies (node:-builtins).

import { readdirSync, readFileSync, existsSync } from 'node:fs';
import { join, relative, resolve } from 'node:path';
import { fileURLToPath } from 'node:url';
import { parseFrontmatter } from '../lib/frontmatter.mjs';
import { TYPE_VOCAB } from '../lib/okf-vocab.mjs';
import { resolveBundleLink } from '../lib/okf-links.mjs';

const RECOMMENDED = ['resource', 'title', 'description', 'timestamp'];

// Lenke-deteksjon for --strict-ingest: hvert lenke-maal maa vaere en trygg,
// on-disk bundle-root-relativ .md (anti-RAG-poison). B3 (A1): ALLE fire
// standard lenkeformer fanges -- inline, referanse-definisjon, autolink og
// raa HTML-anker -- ikke bare inline (de tre siste lakk foer A1).
const MD_LINK_RE = /\[[^\]]*\]\(([^)]+)\)/g;
const REF_DEF_RE = /^[ \t]{0,3}\[[^\]]+\]:[ \t]*(\S+)/gm;
const AUTOLINK_RE = /<([a-z][a-z0-9+.-]*:[^<>\s]*)>/gi;
const HTML_HREF_RE = /<a\b[^>]*\bhref\s*=\s*["']?([^"'\s>]+)/gi;
const LINK_FORMS = [MD_LINK_RE, REF_DEF_RE, AUTOLINK_RE, HTML_HREF_RE];
// En title som selv baerer en markdown-lenke er en injeksjons-vektor -> avvises.
const TITLE_LINK_RE = /\[[^\]]*\]\([^)]*\)/;

// Drop-zone (raa innboks-filer) + skjulte kataloger (.cache osv.) skannes ALDRI:
// raa, ukonverterte og potensielt fiendtlige filer er ikke konsepter.
function isWalkableDir(name) {
  return !name.startsWith('.') && name !== 'innboks';
}

// Alle konsept-filer (.md unntatt index.md) under root, rekursivt.
function walkConcepts(root) {
  const out = [];
  const walk = (dir) => {
    for (const e of readdirSync(dir, { withFileTypes: true })) {
      const p = join(dir, e.name);
      if (e.isDirectory()) {
        if (isWalkableDir(e.name)) walk(p);
      } else if (e.isFile() && e.name.endsWith('.md') && e.name !== 'index.md') {
        out.push(p);
      }
    }
  };
  walk(root);
  return out;
}

// Les rotens okf_version (markdown-tekst i index.md, ikke frontmatter). null hvis fravaerende.
function rootOkfVersion(root) {
  const idx = join(root, 'index.md');
  if (!existsSync(idx)) return null;
  const m = readFileSync(idx, 'utf8').match(/^okf_version:\s*(.+)$/m);
  return m ? m[1].trim() : null;
}

// strictIngest (default AV): paa skrivestien handheves det lukkede vokabularet
// + lenke-allow-listen. Lese-siden (default) forblir tolerant (exit 0/1 kun paa
// manglende type) -- denne porten er aktiv KUN ved ingestion (--strict-ingest).
// files (B2, A1): orkestratoren scoper strict til KUN kjoeringens skrevne filer
// (writeConcepts-returens {concepts, pointers}) -- ALDRI hele roten, som ville
// felt legitimt haandkuratert innhold (lovlig out-of-vocab/eksterne lenker paa
// lese-siden). Uten files: full-root-walk som foer (CLI/lese-siden).
export function checkBundle(root, { strictIngest = false, files } = {}) {
  const concepts = Array.isArray(files)
    ? files.map((f) => resolve(root, f))
    : walkConcepts(root);
  const missingType = [];
  const warnings = [];
  const strictErrors = [];
  for (const f of concepts) {
    const raw = readFileSync(f, 'utf8');
    const { get } = parseFrontmatter(raw);
    const rel = relative(root, f);
    const type = get('type');
    if (!type) {
      missingType.push(rel);
      continue;
    }
    if (strictIngest) {
      // 1. type maa vaere i det lukkede ingestion-vokabularet.
      if (!TYPE_VOCAB.includes(type)) {
        strictErrors.push(`${rel}: type «${type}» utenfor ingestion-vokabular`);
      }
      // 2. en title som baerer en markdown-lenke er en injeksjons-vektor.
      const title = get('title');
      if (title && TITLE_LINK_RE.test(title)) {
        strictErrors.push(`${rel}: lenke-baerende title «${title}»`);
      }
      // 3. hvert lenke-maal (alle fire lenkeformer, B3) maa resolvere til en
      //    trygg, on-disk bundle-fil.
      for (const re of LINK_FORMS) {
        for (const m of raw.matchAll(re)) {
          const target = m[1];
          const resolved = resolveBundleLink(target, root);
          if (!resolved) strictErrors.push(`${rel}: utrygg lenke ${target}`);
          else if (!existsSync(resolved)) strictErrors.push(`${rel}: dangling lenke ${target}`);
        }
      }
    }
    for (const field of RECOMMENDED) {
      if (!get(field)) warnings.push(`${rel}: mangler anbefalt felt «${field}»`);
    }
  }
  return {
    scanned: concepts.length,
    missingType,
    warnings,
    strictErrors,
    strictIngest,
    okfVersion: rootOkfVersion(root),
  };
}

// --- CLI ---
const isMain = process.argv[1]
  && fileURLToPath(import.meta.url) === process.argv[1];
if (isMain) {
  const args = process.argv.slice(2);
  const strictIngest = args.includes('--strict-ingest');
  const root = args.find((a) => !a.startsWith('--'));
  if (!root) {
    process.stderr.write('Bruk: node okf-check.mjs <bundle-rot> [--strict-ingest]\n');
    process.exit(2);
  }
  if (!existsSync(root)) {
    process.stderr.write(`Bundle-rot finnes ikke: ${root}\n`);
    process.exit(2);
  }
  const r = checkBundle(root, { strictIngest });
  const out = [];
  out.push(`OKF-sjekk: ${root}${strictIngest ? ' (strict-ingest)' : ''}`);
  out.push(`  Konsept-filer skannet: ${r.scanned}`);
  out.push(`  ${r.missingType.length} filer uten type:`);
  for (const f of r.missingType) out.push(`    - ${f}`);
  if (strictIngest) {
    out.push(`  ${r.strictErrors.length} strict-ingest-feil:`);
    for (const e of r.strictErrors) out.push(`    x ${e}`);
  }
  out.push(`  okf_version: ${r.okfVersion || 'MANGLER (rot-index uten okf_version)'}`);
  out.push(`  Advarsler (anbefalte felt): ${r.warnings.length}`);
  for (const w of r.warnings) out.push(`    ! ${w}`);
  const failed = r.missingType.length > 0 || (strictIngest && r.strictErrors.length > 0);
  let verdict;
  if (!failed) verdict = 'OK: gyldig OKF-bundle';
  else if (r.missingType.length > 0) verdict = `FEIL: ${r.missingType.length} fil(er) mangler type:`;
  else verdict = `FEIL: ${r.strictErrors.length} strict-ingest-feil`;
  out.push(verdict);
  process.stdout.write(`${out.join('\n')}\n`);
  process.exit(failed ? 1 : 0);
}