ms-ai-architect/tests/kb-update/test-strip-stale-verified-pipe.test.mjs

// tests/kb-update/test-strip-stale-verified-pipe.test.mjs
// TDD for RX-KB1 — strip the stale plain `| Verified: MCP <date>` pipe-tail from the
// **Last updated:** line across the 87 reference files that carry it in the top-500-byte
// header window. This is the same poison class as the 14 bold `**Verified:** MCP` labels
// Spor 1 already removed: a plain-text "Verified" claim the judge never made, invisible to
// the bold-only contract stack (kb-headers.mjs / audit RE_VERIFIED). "Strip" = remove ONLY
// the ` | Verified: …` tail, preserving `**Last updated:** <date>` byte-exact; the citation,
// the date, and the whole body stay untouched. Value-preserving, idempotent, one line edited
// in place (never removed).
//
// The two dialect variants in the real corpus (ground truth 2026-07-16):
//   85× `**Last updated:** <date> | Verified: MCP <date>`
//    2× `**Last updated:** <date> | Verified: <date>`   (no "MCP")

import { test } from 'node:test';
import assert from 'node:assert/strict';
import { stripVerifiedPipe, MANIFEST } from '../../scripts/kb-update/strip-stale-verified-pipe.mjs';

// A reference-shaped fixture: bold header with the plain Verified pipe-tail, then a body that
// itself mentions "Verified:" and even "| Verified:" in prose (must never be touched).
const FIXTURE = [
  '# Azure AI Search — RAG Setup',
  '**Type:** reference',
  '**Last updated:** 2026-06-19 | Verified: MCP 2026-06-19',
  '**Category:** RAG',
  '',
  '---',
  '',
  '## Innhold',
  '',
  'A body line mentioning Verified: yesterday and a | Verified: pipe — must be left alone.',
  '',
].join('\n');

// ── stripVerifiedPipe (pure) ─────────────────────────────────────────────────
test('stripVerifiedPipe: removes the `| Verified: MCP <date>` tail, keeps Last updated byte-exact', () => {
  const out = stripVerifiedPipe(FIXTURE);
  const header = out.split('\n---')[0];
  assert.ok(header.includes('**Last updated:** 2026-06-19'), 'Last updated + date preserved');
  assert.ok(!/Verified:/.test(header), 'no Verified label remains in the header');
  assert.ok(!/\|\s*$/.test(header.split('\n')[2]), 'no dangling pipe/whitespace after strip');
  assert.equal(header.split('\n')[2], '**Last updated:** 2026-06-19', 'exact stripped line');
});

test('stripVerifiedPipe: strips the non-MCP `| Verified: <date>` variant too', () => {
  const nonMcp = FIXTURE.replace('| Verified: MCP 2026-06-19', '| Verified: 2026-06-18');
  const out = stripVerifiedPipe(nonMcp);
  const header = out.split('\n---')[0];
  assert.equal(header.split('\n')[2], '**Last updated:** 2026-06-19', 'non-MCP tail stripped, date kept');
  assert.ok(!/Verified:/.test(header), 'no Verified label remains');
});

test('stripVerifiedPipe: one line edited in place — total line count unchanged', () => {
  const out = stripVerifiedPipe(FIXTURE);
  assert.equal(out.split('\n').length, FIXTURE.split('\n').length, 'no line added or removed');
});

test('stripVerifiedPipe: body is byte-identical (only the header line changes)', () => {
  const out = stripVerifiedPipe(FIXTURE);
  const bodyOf = (c) => c.slice(c.indexOf('\n---'));
  assert.equal(bodyOf(out), bodyOf(FIXTURE), 'body untouched');
});

test('stripVerifiedPipe: never touches a "Verified:" or "| Verified:" in the body', () => {
  const out = stripVerifiedPipe(FIXTURE);
  assert.ok(out.includes('mentioning Verified: yesterday and a | Verified: pipe'), 'body prose intact');
});

test('stripVerifiedPipe: idempotent — a second pass is a no-op', () => {
  const once = stripVerifiedPipe(FIXTURE);
  assert.equal(stripVerifiedPipe(once), once);
});

test('stripVerifiedPipe: no pipe-tail present → unchanged (clean Last updated)', () => {
  const clean = FIXTURE.replace(' | Verified: MCP 2026-06-19', '');
  assert.equal(stripVerifiedPipe(clean), clean);
});

test('stripVerifiedPipe: leaves the citation date byte-exact even if it differs from the Verified date', () => {
  // Last updated 2026-06-24, Verified MCP 2026-06 (the drift the review flagged) — the strip
  // keeps the authored Last updated date and simply drops the unverifiable Verified claim.
  const drift = FIXTURE.replace('**Last updated:** 2026-06-19 | Verified: MCP 2026-06-19',
                                '**Last updated:** 2026-06-24 | Verified: MCP 2026-06');
  const out = stripVerifiedPipe(drift);
  assert.equal(out.split('\n')[2], '**Last updated:** 2026-06-24');
});

// ── MANIFEST (frozen ground truth 2026-07-16) ────────────────────────────────
test('MANIFEST: exactly the 87 plain-Verified files, no duplicates', () => {
  assert.equal(MANIFEST.length, 87, 'frozen count');
  assert.equal(new Set(MANIFEST).size, 87, 'no duplicates');
});

test('MANIFEST: per-skill breakdown 18/45/8/16 (advisor/engineering/governance/security)', () => {
  const bySkill = (skill) => MANIFEST.filter((p) => p.startsWith(`skills/${skill}/`)).length;
  assert.equal(bySkill('ms-ai-advisor'), 18);
  assert.equal(bySkill('ms-ai-engineering'), 45);
  assert.equal(bySkill('ms-ai-governance'), 8);
  assert.equal(bySkill('ms-ai-security'), 16);
});

test('MANIFEST: every entry is a plugin-relative references .md path', () => {
  for (const p of MANIFEST) {
    assert.match(p, /^skills\/ms-ai-[a-z]+\/references\/.+\.md$/, `bad path: ${p}`);
  }
});

test('MANIFEST: sorted (deterministic, diff-stable)', () => {
  const sorted = [...MANIFEST].sort();
  assert.deepEqual(MANIFEST, sorted, 'manifest must be sorted');
});