@opengsd/gsd-core 1.9.1 → 1.11.0

package/scripts/lint-example-parser-parity.cjs

@@ -0,0 +1,395 @@
+#!/usr/bin/env node
+'use strict';
+
+/**
+ * lint-example-parser-parity.cjs — asserts examples/dynamic-context-management/
+ * context-predicates.cjs (reference prototype, ADR-1671) and production's
+ * src/context-predicates.cts (compiled to gsd-core/bin/lib/context-predicates.cjs)
+ * agree on parsing behavior, and that the example's own committed
+ * CONTEXT-INDEX.json has not silently drifted from a fresh parse.
+ *
+ * Lives OUTSIDE tests/ deliberately. ADR-1671
+ * (docs/adr/1671-dynamic-context-management-platform.md:102) places
+ * examples/dynamic-context-management/ outside FOUR surfaces: the build
+ * (src/ -> bin/lib/), the npm package files[], the installer, and the CI test
+ * suite (tests/). A lint script wired into `npm run lint:ci` is none of
+ * those four — it does not compile, package, install, or test-suite-import
+ * the example; it only READS both modules from a repo-root script and
+ * asserts they agree, the same shape as every other scripts/lint-*.cjs in
+ * this repo that reads source it does not own (see e.g.
+ * lint-compiled-artifact-sync.cjs).
+ *
+ * Two parity surfaces:
+ *   1. production vs. example, parsed against the real repo-root CONTEXT.md:
+ *      predicate count, class map, duplicate-id set, the full set of
+ *      (id, value) pairs, and a representative accept/reject id-shape table
+ *      (including the embedded-CR-in-value case production already rejected
+ *      and the example silently accepted before this fix).
+ *   2. the example's OWN committed CONTEXT-INDEX.json vs. a fresh parse by
+ *      the example's own parser of the real CONTEXT.md — LINE-NUMBER-
+ *      INDEPENDENT (see below). This guards the same #2944 merge-race
+ *      staleness class this PR fixed for docs/CONTEXT-INDEX.json, but for the
+ *      example's own copy, which nothing previously guarded at all.
+ *
+ * Why `line` is excluded from surface 2 (ADR-1671 open question 4): the
+ * example's CONTEXT-INDEX.json bakes source line numbers into every
+ * predicate entry. Any unrelated CONTEXT.md line-shift (e.g. inserting a
+ * sentence above the predicates) re-drifts that artifact even though every
+ * predicate id/value/class is byte-identical. Wiring the example's own
+ * `gen-context-index.cjs --check` into lint:ci — which DOES compare line
+ * numbers — would make CI routinely red for reasons unrelated to predicate
+ * integrity. This lint instead derives the LINE-INDEPENDENT facts from the
+ * committed index (count, class map, duplicate-id set, (id,value) pairs) and
+ * compares those to a fresh parse: genuine content drift (the #2944
+ * merge-race defect class) is still caught; a pure line-shift is not.
+ *
+ * `--context-path` / `--example-index-path` override the two hardcoded paths
+ * (mirrors scripts/gen-context-index.cjs) so this script's non-vacuousness
+ * can be proven against a throwaway temp fixture without ever touching the
+ * real committed artifacts.
+ */
+
+const fs = require('node:fs');
+const path = require('node:path');
+
+const { ExitError, runMain } = require('./lib/cli-exit.cjs');
+
+const ROOT = path.resolve(__dirname, '..');
+const PROD_PREDICATES_PATH = path.join(ROOT, 'gsd-core', 'bin', 'lib', 'context-predicates.cjs');
+const EXAMPLE_DIR = path.join(ROOT, 'examples', 'dynamic-context-management');
+const EXAMPLE_PREDICATES_PATH = path.join(EXAMPLE_DIR, 'context-predicates.cjs');
+const DEFAULT_EXAMPLE_INDEX_PATH = path.join(EXAMPLE_DIR, 'CONTEXT-INDEX.json');
+const DEFAULT_CONTEXT_PATH = path.join(ROOT, 'CONTEXT.md');
+
+// ─── Loaders ──────────────────────────────────────────────────────────────────
+
+/**
+ * Load the compiled production predicates library. Throws a clean ExitError
+ * (never a bare MODULE_NOT_FOUND stack) naming the remedy when it is missing.
+ */
+function loadProdPredicates() {
+  try {
+    delete require.cache[require.resolve(PROD_PREDICATES_PATH)];
+    return require(PROD_PREDICATES_PATH);
+  } catch (err) {
+    throw new ExitError(
+      1,
+      `Cannot load ${path.relative(ROOT, PROD_PREDICATES_PATH)}: ${err && err.message}\n` +
+        'Run:\n  npm run build:lib\n',
+    );
+  }
+}
+
+/** Load the example's self-contained predicates module (no build step). */
+function loadExamplePredicates() {
+  delete require.cache[require.resolve(EXAMPLE_PREDICATES_PATH)];
+  return require(EXAMPLE_PREDICATES_PATH);
+}
+
+/**
+ * Read a CONTEXT.md-shaped markdown file. Throws a clean ExitError naming the
+ * path when it is missing or unreadable.
+ *
+ * @param {string} contextPath
+ * @returns {string}
+ */
+function readContextMarkdown(contextPath) {
+  try {
+    return fs.readFileSync(contextPath, 'utf8');
+  } catch (err) {
+    throw new ExitError(1, `Cannot read ${path.relative(ROOT, contextPath)}: ${err && err.message}`);
+  }
+}
+
+// ─── Diffing helpers ────────────────────────────────────────────────────────
+
+/**
+ * Diff two predicate-shaped arrays by id, comparing ONLY `klass`/`value`
+ * (never `line`/`section` — see the module doc comment's "Why `line` is
+ * excluded"). Returns human-readable divergence strings naming the exact
+ * predicate id(s), so a failure reads as an actionable list, never "objects
+ * differ".
+ *
+ * @param {Array<{id:string,klass:string,value:string}>} leftPredicates
+ * @param {Array<{id:string,klass:string,value:string}>} rightPredicates
+ * @param {string} leftLabel
+ * @param {string} rightLabel
+ * @returns {string[]}
+ */
+function diffPredicatesById(leftPredicates, rightPredicates, leftLabel, rightLabel) {
+  const leftMap = new Map(leftPredicates.map((p) => [p.id, p]));
+  const rightMap = new Map(rightPredicates.map((p) => [p.id, p]));
+  const allIds = new Set([...leftMap.keys(), ...rightMap.keys()]);
+  const diffs = [];
+  for (const id of Array.from(allIds).sort()) {
+    const l = leftMap.get(id);
+    const r = rightMap.get(id);
+    if (l && !r) {
+      diffs.push(`${id}: present in ${leftLabel} but absent from ${rightLabel}`);
+    } else if (!l && r) {
+      diffs.push(`${id}: present in ${rightLabel} but absent from ${leftLabel}`);
+    } else if (l.value !== r.value) {
+      diffs.push(
+        `${id}: value diverged (${leftLabel}=${JSON.stringify(l.value)}, ${rightLabel}=${JSON.stringify(r.value)})`,
+      );
+    } else if (l.klass !== r.klass) {
+      diffs.push(
+        `${id}: klass diverged (${leftLabel}=${JSON.stringify(l.klass)}, ${rightLabel}=${JSON.stringify(r.klass)})`,
+      );
+    }
+  }
+  return diffs;
+}
+
+/** Sorted, deduplicated id list from a `duplicates` array (either `{id,count}` or `{id,lines}` shape). */
+function duplicateIds(duplicates) {
+  return duplicates.map((d) => d.id).sort();
+}
+
+/** Build a single-line `` `id=value` `` markdown fixture wrapping one candidate id=value pair. */
+function backtickLine(idEqualsValue) {
+  return '`' + idEqualsValue + '`';
+}
+
+// Representative accept/reject id-shape table both parsers must agree on —
+// this is what PROVES the parity claim rather than merely asserting it.
+// FINDING 4 (ADR-1671 PR review): the embedded-CR-in-value case below was the
+// one real divergence the earlier parity claim missed — production rejected
+// it, the example silently accepted it. Both now reject it (ported fix).
+const ID_SHAPE_CASES = [
+  { idEqualsValue: 'A=1', expectAccept: true },
+  { idEqualsValue: 'FOO=x', expectAccept: true },
+  { idEqualsValue: 'PRED.k320.rule=x', expectAccept: true },
+  { idEqualsValue: 'RELEASE-NOTES.x=y', expectAccept: true },
+  { idEqualsValue: 'A..b=1', expectAccept: false },
+  { idEqualsValue: 'foo.bar=x', expectAccept: false },
+  { idEqualsValue: 'FOO BAR=x', expectAccept: false },
+  { idEqualsValue: '=v', expectAccept: false },
+  { idEqualsValue: 'ID', expectAccept: false },
+];
+
+/**
+ * The embedded-CR case needs its own fixture, not the plain
+ * `` `ID=value` `` shape backtickLine() builds: a lone CR with no following
+ * LF never reaches a closing backtick at all (a separate, documented parser
+ * limit — see src/context-predicates.cts's module doc comment), so the CR
+ * must sit INSIDE an otherwise well-formed, LF-terminated backtick line.
+ */
+const CR_VALUE_MARKDOWN = '`ID=ab\rcd`\n';
+
+// ─── Parity checks (surface 1: production vs. example) ────────────────────────
+
+/**
+ * @param {object} prodPredicates
+ * @param {object} examplePredicates
+ * @param {string} markdown
+ * @returns {string[]} diagnostics; empty when the two parsers fully agree
+ */
+function checkProdExampleParity(prodPredicates, examplePredicates, markdown) {
+  const diagnostics = [];
+
+  const prodParsed = prodPredicates.parsePredicates(markdown);
+  const exampleParsed = examplePredicates.parsePredicates(markdown);
+  const prodIndex = prodPredicates.buildIndex(prodParsed.predicates);
+  const exampleIndex = examplePredicates.buildIndex(exampleParsed.predicates);
+
+  if (exampleIndex.count !== prodIndex.count) {
+    diagnostics.push(`predicate count diverged: example=${exampleIndex.count} production=${prodIndex.count}`);
+  }
+
+  const prodClassKeys = Object.keys(prodIndex.classes).sort();
+  const exampleClassKeys = Object.keys(exampleIndex.classes).sort();
+  if (JSON.stringify(exampleClassKeys) !== JSON.stringify(prodClassKeys)) {
+    diagnostics.push(
+      `class set diverged: example=${JSON.stringify(exampleClassKeys)} production=${JSON.stringify(prodClassKeys)}`,
+    );
+  } else if (JSON.stringify(exampleIndex.classes) !== JSON.stringify(prodIndex.classes)) {
+    diagnostics.push(
+      `per-class predicate counts diverged: example=${JSON.stringify(exampleIndex.classes)} ` +
+        `production=${JSON.stringify(prodIndex.classes)}`,
+    );
+  }
+
+  const prodDupIds = duplicateIds(prodIndex.duplicates);
+  const exampleDupIds = duplicateIds(exampleIndex.duplicates);
+  if (JSON.stringify(exampleDupIds) !== JSON.stringify(prodDupIds)) {
+    diagnostics.push(
+      `duplicate-id set diverged: example=${JSON.stringify(exampleDupIds)} production=${JSON.stringify(prodDupIds)}`,
+    );
+  }
+
+  diagnostics.push(...diffPredicatesById(exampleIndex.predicates, prodIndex.predicates, 'example', 'production'));
+
+  for (const { idEqualsValue, expectAccept } of ID_SHAPE_CASES) {
+    const line = backtickLine(idEqualsValue);
+    const exampleCount = examplePredicates.parsePredicates(line).predicates.length;
+    const prodCount = prodPredicates.parsePredicates(line).predicates.length;
+    if (exampleCount !== prodCount) {
+      diagnostics.push(
+        `id-shape verdict diverged for ${JSON.stringify(idEqualsValue)}: ` +
+          `example found ${exampleCount} predicate(s), production found ${prodCount}`,
+      );
+      continue;
+    }
+    const actualAccept = prodCount === 1;
+    if (actualAccept !== expectAccept) {
+      diagnostics.push(
+        `id-shape verdict wrong for ${JSON.stringify(idEqualsValue)}: expected ` +
+          `${expectAccept ? 'ACCEPT' : 'REJECT'}, both modules agreed on ${actualAccept ? 'ACCEPT' : 'REJECT'} instead`,
+      );
+    }
+  }
+
+  const exampleCrCount = examplePredicates.parsePredicates(CR_VALUE_MARKDOWN).predicates.length;
+  const prodCrCount = prodPredicates.parsePredicates(CR_VALUE_MARKDOWN).predicates.length;
+  if (exampleCrCount !== prodCrCount) {
+    diagnostics.push(
+      `id-shape verdict diverged for embedded-CR-in-value: example found ${exampleCrCount} predicate(s), ` +
+        `production found ${prodCrCount}`,
+    );
+  } else if (prodCrCount !== 0) {
+    diagnostics.push(
+      `embedded-CR-in-value must be REJECTED by both parsers; both instead ACCEPTED it (${prodCrCount} predicate(s))`,
+    );
+  }
+
+  return diagnostics;
+}
+
+// ─── Parity check (surface 2: example's committed index vs. fresh parse) ──────
+
+/** Strip a predicate down to the line-number-independent fields used for comparison. */
+function toLineIndependent(p) {
+  return { id: p.id, klass: p.klass, value: p.value };
+}
+
+/**
+ * @param {object} examplePredicates
+ * @param {string} markdown
+ * @param {string} exampleIndexPath
+ * @returns {string[]} diagnostics; empty when the committed index and a fresh
+ *   parse agree on every LINE-NUMBER-INDEPENDENT fact
+ */
+function checkExampleIndexParity(examplePredicates, markdown, exampleIndexPath) {
+  const diagnostics = [];
+
+  let committed;
+  try {
+    committed = JSON.parse(fs.readFileSync(exampleIndexPath, 'utf8'));
+  } catch (err) {
+    diagnostics.push(`cannot read/parse ${exampleIndexPath}: ${err && err.message}`);
+    return diagnostics;
+  }
+  if (!committed || !Array.isArray(committed.predicates)) {
+    diagnostics.push(`${exampleIndexPath} does not have the expected ContextIndex shape`);
+    return diagnostics;
+  }
+
+  const { predicates } = examplePredicates.parsePredicates(markdown);
+  const fresh = examplePredicates.buildIndex(predicates);
+
+  if (committed.count !== fresh.count) {
+    diagnostics.push(`example index predicate count diverged: committed=${committed.count} fresh=${fresh.count}`);
+  }
+  if (JSON.stringify(committed.classes) !== JSON.stringify(fresh.classes)) {
+    diagnostics.push(
+      `example index class-count map diverged: committed=${JSON.stringify(committed.classes)} ` +
+        `fresh=${JSON.stringify(fresh.classes)}`,
+    );
+  }
+
+  const committedDupIds = duplicateIds(committed.duplicates || []);
+  const freshDupIds = duplicateIds(fresh.duplicates);
+  if (JSON.stringify(committedDupIds) !== JSON.stringify(freshDupIds)) {
+    diagnostics.push(
+      `example index duplicate-id set diverged: committed=${JSON.stringify(committedDupIds)} ` +
+        `fresh=${JSON.stringify(freshDupIds)}`,
+    );
+  }
+
+  const committedIndependent = committed.predicates.map(toLineIndependent);
+  const freshIndependent = fresh.predicates.map(toLineIndependent);
+  const diffs = diffPredicatesById(committedIndependent, freshIndependent, 'committed example index', 'fresh parse');
+  diagnostics.push(...diffs.map((d) => `example index: ${d}`));
+
+  return diagnostics;
+}
+
+// ─── Argument parsing ─────────────────────────────────────────────────────────
+
+/**
+ * @param {string[]} argv - process.argv.slice(2)
+ * @returns {{ contextPath: string, exampleIndexPath: string }}
+ */
+function parseArgs(argv) {
+  const opts = { contextPath: DEFAULT_CONTEXT_PATH, exampleIndexPath: DEFAULT_EXAMPLE_INDEX_PATH };
+  for (let i = 0; i < argv.length; i++) {
+    const arg = argv[i];
+    if (arg === '--context-path') {
+      opts.contextPath = path.resolve(argv[++i] ?? '');
+    } else if (arg === '--example-index-path') {
+      opts.exampleIndexPath = path.resolve(argv[++i] ?? '');
+    } else {
+      throw new ExitError(
+        1,
+        `Unknown argument: ${arg}\n` +
+          'Usage: lint-example-parser-parity.cjs [--context-path <path>] [--example-index-path <path>]\n',
+      );
+    }
+  }
+  return opts;
+}
+
+// ─── Main ─────────────────────────────────────────────────────────────────────
+
+function main() {
+  const opts = parseArgs(process.argv.slice(2));
+
+  const prodPredicates = loadProdPredicates();
+  const examplePredicates = loadExamplePredicates();
+  const markdown = readContextMarkdown(opts.contextPath);
+
+  const diagnostics = [
+    ...checkProdExampleParity(prodPredicates, examplePredicates, markdown),
+    ...checkExampleIndexParity(examplePredicates, markdown, opts.exampleIndexPath),
+  ];
+
+  if (diagnostics.length > 0) {
+    process.stderr.write(
+      `\nERROR lint-example-parser-parity: ${diagnostics.length} divergence(s) found\n\n`,
+    );
+    for (const d of diagnostics) process.stderr.write(`  - ${d}\n`);
+    process.stderr.write(
+      '\nexamples/dynamic-context-management/context-predicates.cjs and ' +
+        'src/context-predicates.cts (gsd-core/bin/lib/context-predicates.cjs) must agree; ' +
+        'the committed examples/dynamic-context-management/CONTEXT-INDEX.json must match a ' +
+        'fresh parse of CONTEXT.md on every line-number-independent fact (run:\n' +
+        '  node examples/dynamic-context-management/gen-context-index.cjs --write\n' +
+        ').\n',
+    );
+    throw new ExitError(1);
+  }
+
+  process.stdout.write(
+    'ok lint-example-parser-parity: production and example parsers agree on the real CONTEXT.md ' +
+      `(count/classes/duplicates/(id,value) pairs), on ${ID_SHAPE_CASES.length + 1} representative ` +
+      'accept/reject id-shape cases, and the committed example index matches a fresh parse ' +
+      '(line-number-independent)\n',
+  );
+}
+
+module.exports = {
+  loadProdPredicates,
+  loadExamplePredicates,
+  readContextMarkdown,
+  diffPredicatesById,
+  checkProdExampleParity,
+  checkExampleIndexParity,
+  parseArgs,
+  ID_SHAPE_CASES,
+  CR_VALUE_MARKDOWN,
+};
+
+if (require.main === module) {
+  runMain(main);
+}

package/scripts/lint-frontmatter-scalar-broad-grep.cjs

@@ -0,0 +1,237 @@
+#!/usr/bin/env node
+'use strict';
+
+/**
+ * lint-frontmatter-scalar-broad-grep.cjs — DEFECT.FRONTMATTER-SCALAR-BROAD-GREP
+ * (CONTEXT.md).
+ *
+ * ## Why
+ *
+ * A YAML-frontmatter scalar (e.g. VERIFICATION.md `status:`) read with
+ * `grep "^key:"` over the WHOLE markdown report instead of the frontmatter
+ * block returns extra matches whenever a `key:` line also appears in the
+ * body (a code block, a copied artifact, an example). Piped into
+ * `cut`/`tr`, those extra matches concatenate into a value that matches no
+ * expected token, silently misrouting a valid state (#586/PR #650:
+ * `grep "^status:"` also matched body `status:` lines, yielding
+ * `passed+gaps_found+human_needed` instead of `passed` and blocking a
+ * passed phase).
+ *
+ * The fix-forward is to scope the grep to the leading frontmatter block and
+ * take only the first match:
+ *   sed -n '/^---$/,/^---$/p' "$f" | grep -m1 "^<key>:" | cut -d: -f2 | tr -d ' '
+ *
+ * ## What this scans
+ *
+ * Every fenced ```bash / ```sh code block in `gsd-core/workflows/*.md`,
+ * `agents/*.md`, and `commands/**\/*.md`. Within each block, flags a
+ * `grep "^key:"` / `grep '^key:'` invocation that:
+ *   - is NOT preceded (earlier in the SAME block) by a frontmatter-scoping
+ *     idiom (`sed -n '/^---$/,/^---$/p'`, a JS `/^---\n([\s\S]*?)\n---/`
+ *     extraction, or an equivalent range over the `---` delimiter), AND
+ *   - does NOT carry a `-m1` (or `-m 1`) flag, and is NOT immediately piped
+ *     into `head -1`/`head -n 1` (frontmatter always precedes the body in
+ *     these generated reports, so `head -1` on the whole file is the same
+ *     single-match guarantee as `-m1`), AND
+ *   - is used for exact-token comparison: piped (same line) into
+ *     `cut`/`tr`, or captured into a shell variable that is later compared
+ *     via `==`/`case` elsewhere in the same block.
+ *
+ * ## False-positive risk (moderate-to-high, per audit)
+ *
+ * Some `grep "^key:"` uses are intentionally whole-body (scanning multiple
+ * report files at once, not one frontmatter block) and are not a bug. Add
+ * `# lint-allow: frontmatter-scalar-broad-grep — <reason>` on the same line
+ * (or the line immediately above) to suppress a specific invocation.
+ */
+
+const fs = require('fs');
+const path = require('path');
+const { ExitError, runMain } = require('./lib/cli-exit.cjs');
+
+const ROOT = path.join(__dirname, '..');
+const DEFAULT_ROOTS = ['gsd-core/workflows', 'agents', 'commands'];
+
+const FENCE_RE = /^```(bash|sh)\s*$/;
+const FENCE_END_RE = /^```\s*$/;
+
+// A `grep "^key:"` / `grep '^key:'` invocation. Captures the key name and the
+// full option string preceding the pattern (so callers can check for -m1).
+const GREP_KEY_RE = /grep\s+((?:-\S+\s+)*)(["'])\^([A-Za-z_][\w-]*):\2/;
+
+// A frontmatter-scoping idiom: a delimiter-range extraction anchored on the
+// `---` frontmatter fence, opened by `^---` (sed/awk `/^---$/,/^---$/p`, or a
+// JS regex like `/^---\n([\s\S]*?)\n---/`) and closed by a second `---`
+// within a short window. Matches both idioms without caring which language
+// wrote the delimiter.
+const FRONTMATTER_SCOPE_RE = /\^---[\s\S]{0,300}?---/;
+
+const ALLOW_RE = /#\s*lint-allow:\s*frontmatter-scalar-broad-grep/;
+
+// `| head -1` / `| head -n 1` immediately after the grep is functionally
+// equivalent to `-m1` for this check: frontmatter always precedes the body
+// in these generated reports, so the first grep match is always the
+// frontmatter's, and `head -1` discards every later (body) match exactly
+// like `-m1` would.
+function hasSingleMatchGuard(line, optionString) {
+  if (/(^|\s)-m\s*1(\s|$)/.test(optionString) || /(^|\s)--max-count[= ]1(\s|$)/.test(optionString)) return true;
+  return /\|\s*head\s+(-1|-n\s*1)\b/.test(line);
+}
+
+function isSuppressed(lines, idx) {
+  if (ALLOW_RE.test(lines[idx])) return true;
+  if (idx > 0 && ALLOW_RE.test(lines[idx - 1])) return true;
+  return false;
+}
+
+/**
+ * Extract fenced ```bash/```sh code blocks from markdown text.
+ * @param {string} text
+ * @returns {{ startLine: number, lines: string[] }[]}
+ */
+function extractBashBlocks(text) {
+  const allLines = text.split(/\r?\n/);
+  const blocks = [];
+  let inBlock = false;
+  let blockLines = [];
+  let blockStart = 0;
+  for (let i = 0; i < allLines.length; i += 1) {
+    const line = allLines[i];
+    if (!inBlock && FENCE_RE.test(line.trim())) {
+      inBlock = true;
+      blockLines = [];
+      blockStart = i + 2; // first line INSIDE the block is 1-indexed i+2
+      continue;
+    }
+    if (inBlock && FENCE_END_RE.test(line.trim())) {
+      blocks.push({ startLine: blockStart, lines: blockLines });
+      inBlock = false;
+      continue;
+    }
+    if (inBlock) blockLines.push(line);
+  }
+  return blocks;
+}
+
+/**
+ * Pure: find every un-scoped, token-comparison `grep "^key:"` invocation in a
+ * single fenced bash/sh block's lines. Returns `{ line, key, snippet }[]`
+ * (line numbers relative to the block's startLine, already offset by caller).
+ * @param {string[]} lines
+ * @returns {{ lineIndex: number, key: string, snippet: string }[]}
+ */
+function findBroadGrepsInBlock(lines) {
+  const findings = [];
+  // Variables assigned from a grep-key capture on this block, so a later
+  // `==`/`case` use of that variable (without an intervening scope/-m1) also
+  // counts as "used for exact-token comparison".
+  const capturedVars = new Set();
+  let scopeSeenAt = -1;
+
+  for (let i = 0; i < lines.length; i += 1) {
+    const line = lines[i];
+
+    if (FRONTMATTER_SCOPE_RE.test(line)) {
+      scopeSeenAt = i;
+    }
+
+    const m = line.match(GREP_KEY_RE);
+    if (!m) continue;
+    const [, options, , key] = m;
+    if (hasSingleMatchGuard(line, options)) continue;
+    if (isSuppressed(lines, i)) continue;
+    // Scoping must appear strictly before this grep line in the same block.
+    const scoped = scopeSeenAt !== -1 && scopeSeenAt <= i;
+    if (scoped) continue;
+
+    const pipedToTokenTool = /\|\s*(cut|tr)\b/.test(line);
+    const assignMatch = line.match(/^\s*(?:export\s+)?([A-Za-z_][\w]*)=\$\(/);
+    if (assignMatch) capturedVars.add(assignMatch[1]);
+
+    let comparedLater = false;
+    if (assignMatch) {
+      const varName = assignMatch[1];
+      for (let j = i + 1; j < lines.length; j += 1) {
+        if (
+          new RegExp(`\\$\\{?${varName}\\}?"?\\s*(==|!=)`).test(lines[j])
+          || new RegExp(`case\\s+"?\\$\\{?${varName}\\}?"?\\s+in`).test(lines[j])
+        ) {
+          comparedLater = true;
+          break;
+        }
+      }
+    }
+
+    if (pipedToTokenTool || comparedLater) {
+      findings.push({ lineIndex: i, key, snippet: line.trim() });
+    }
+  }
+  return findings;
+}
+
+function walkMarkdown(dir) {
+  const out = [];
+  let entries;
+  try {
+    entries = fs.readdirSync(dir, { withFileTypes: true });
+  } catch {
+    return out; // a missing root is not an error — some surfaces are optional
+  }
+  for (const entry of entries) {
+    const full = path.join(dir, entry.name);
+    if (entry.isDirectory()) out.push(...walkMarkdown(full));
+    else if (entry.isFile() && entry.name.endsWith('.md')) out.push(full);
+  }
+  return out;
+}
+
+/**
+ * Scan the given roots (repo-relative) for un-scoped frontmatter-scalar
+ * broad-greps.
+ * @param {string[]} roots
+ * @returns {{ file: string, line: number, key: string, snippet: string }[]}
+ */
+function scan(roots = DEFAULT_ROOTS) {
+  const offenders = [];
+  for (const rel of roots) {
+    const abs = path.isAbsolute(rel) ? rel : path.join(ROOT, rel);
+    for (const file of walkMarkdown(abs)) {
+      const blocks = extractBashBlocks(fs.readFileSync(file, 'utf8'));
+      for (const block of blocks) {
+        for (const finding of findBroadGrepsInBlock(block.lines)) {
+          offenders.push({
+            file: path.relative(ROOT, file),
+            line: block.startLine + finding.lineIndex,
+            key: finding.key,
+            snippet: finding.snippet,
+          });
+        }
+      }
+    }
+  }
+  return offenders;
+}
+
+function main() {
+  const rootsEnv = process.env.GSD_LINT_FRONTMATTER_SCALAR_ROOTS;
+  const roots = rootsEnv ? rootsEnv.split(path.delimiter).filter(Boolean) : DEFAULT_ROOTS;
+  const offenders = scan(roots);
+  if (offenders.length > 0) {
+    const detail = offenders.map((o) => `  ${o.file}:${o.line}  ${o.snippet}`).join('\n');
+    throw new ExitError(
+      1,
+      'lint-frontmatter-scalar-broad-grep: `grep "^key:"` over the whole file, compared to an\n'
+        + 'exact token, with no frontmatter scoping and no -m1 (DEFECT.FRONTMATTER-SCALAR-BROAD-GREP).\n'
+        + 'A body line beginning `key:` is enough to break this. Scope to the frontmatter block:\n'
+        + '  sed -n \'/^---$/,/^---$/p\' "$f" | grep -m1 "^<key>:" | cut -d: -f2 | tr -d \' \'\n'
+        + 'or add `# lint-allow: frontmatter-scalar-broad-grep — <reason>` if this is a genuine\n'
+        + 'whole-body scan:\n'
+        + detail,
+    );
+  }
+  console.log(`ok lint-frontmatter-scalar-broad-grep: no un-scoped frontmatter-scalar greps in ${roots.length} root(s)`);
+}
+
+module.exports = { findBroadGrepsInBlock, extractBashBlocks, scan, DEFAULT_ROOTS };
+
+if (require.main === module) runMain(main);