voyage/lib/validators/plan-validator.mjs

// lib/validators/plan-validator.mjs
// Wraps plan-schema (heading shape) + manifest-yaml (per-step Manifest blocks).
// This is the JS equivalent of Phase 5.5 grep checks in planning-orchestrator.
//
// Schema is forward-compatible: unknown top-level frontmatter keys are
// tolerated silently. Strict-key checks are intentionally avoided so new
// optional fields (jf. the source_findings precedent) can be added without
// a plan_version bump.
//
// plan_version 1.8 rules (apply only to plans declaring >= 1.8; 1.7 plans
// validate exactly as before):
//   - every step has `Test first:` naming a File and what it Verifies;
//   - the plan has a "run the thing" step: an `End-to-end:` field with a
//     backticked command (without it a result is "not verified end to end");
//   - the optional step fields `Interface in:`, `Interface out:`,
//     `Review focus:` are never empty when present;
//   - `ceremony:` is small | bounded | architectural (absent = architectural);
//     an architectural plan carries `## Design` (see checkDesign);
//   - every check the run will execute can run (checkRunnable): each
//     `## Verification` criterion, each step's `Verify:` (its first backticked
//     span is the step's test command) and each `End-to-end:` command pass the
//     criteria runner's allowlist, and each `End-to-end:` command also stands
//     under `## Verification`, where Phase 7 runs it.
// Strict mode reports these as errors; --soft downgrades them to warnings.
//
// Every plan version (the fields are new; a plan without them is unchanged):
//   - K10: the optional step field `Regression:` holds ONE command the
//     criteria runner's allowlist runs (first backticked span, no pipe, no
//     `&&`); the executor runs it green before the step's checkpoint. One that
//     can never run is an error (a warning under --soft).
//   - K11: `On failure:` opens with revert | retry | skip | escalate, the words
//     /trekexecute reads; any other opening is a warning.
//   - A step whose `Files:` are all tests (plan-schema `stepTestsOnly`) is the
//     warning PLAN_STEP_TESTS_ONLY: its tests pass at its start point, so it can
//     never be proven red (voyage-lekerepo trial 3, 2026-10-09).

import { readFileSync, existsSync, realpathSync } from 'node:fs';
import { fileURLToPath } from 'node:url';
import {
  sliceSteps, validatePlanHeadings, extractPlanVersion, extractPlanCeremony,
  parseStepFields, findSection, splitSections, findModuleCycle, parseMermaid, fencedBlocks,
  versionAtLeast, PLAN_VERSION_CURRENT, PLAN_VERSIONS_SUPPORTED, CEREMONY_VALUES,
  ON_FAILURE_ACTIONS, onFailureAction, stepTestsOnly,
} from '../parsers/plan-schema.mjs';
import { validateAllManifests } from '../parsers/manifest-yaml.mjs';
import { parseDocument } from '../util/frontmatter.mjs';
import { issue, fail } from '../util/result.mjs';
import { parsePlanVerification, firstCommand, runnableCheck } from '../verification/criteria-runner.mjs';

export const OPTIONAL_STEP_FIELDS = Object.freeze(['Interface in', 'Interface out', 'Review focus']);
export const DESIGN_SUBSECTIONS = Object.freeze([
  'Architecture', 'Data Flow', 'Interfaces', 'Data Models', 'Error Handling',
  'Module Map', 'Architecture Rules', 'Requirements Check',
]);
const BACKTICKED = /`[^`\n]+`/;

function checkSteps(sections) {
  const out = [];
  let endToEnd = 0;
  for (const s of sections) {
    const f = parseStepFields(s.body);
    const tf = f.get('Test first');
    const names = tf && /\bFile:\**[ \t]*`[^`\n]+`/i.test(tf.text) && /\bVerifies:\**[ \t]*\S/i.test(tf.text);
    if (!names) {
      out.push(issue(
        'PLAN_STEP_NO_TEST_FIRST',
        `Step ${s.n}: ${tf ? '`Test first:` does not name a File (backticked path) and what it Verifies' : 'no `Test first:` field'}`,
        'plan_version 1.8 requires `Test first:` on every step, with `File:` and `Verifies:` sub-bullets.',
      ));
    }
    for (const name of OPTIONAL_STEP_FIELDS) {
      if (f.has(name) && f.get(name).text === '') {
        out.push(issue('PLAN_STEP_FIELD_EMPTY', `Step ${s.n}: \`${name}:\` is present but empty`,
          'Fill the field or remove it; it is optional.'));
      }
    }
    if (f.has('End-to-end') && BACKTICKED.test(f.get('End-to-end').text)) endToEnd += 1;
  }
  if (sections.length > 0 && endToEnd === 0) {
    out.push(issue(
      'PLAN_NO_END_TO_END',
      'No "run the thing" step: no step carries `End-to-end:` with a backticked command',
      'Add a step that runs the product the way a user does (`End-to-end: `<command>` → expected: …`). ' +
        'Without it the result is "not verified end to end", never green.',
    ));
  }
  return out;
}

const WAY_IN = 'A check the allowlist has no name for (a pipe, `jq`, `grep`, the product\'s own binary) is ' +
  'wrapped in a script under tests/ and declared as `bash tests/<script>.sh`.';

/**
 * Every check the run will execute can run. Without this a plan passed the
 * validator, the plan gate and three critic rounds with a `git grep` criterion
 * the runner refuses, and Phase 7 felled a run whose steps had all passed
 * (voyage-lekerepo trial, 2026-10-06).
 */
function checkRunnable(text, sections, opts) {
  const out = [];
  const criteria = parsePlanVerification(text);
  for (const c of criteria) {
    const r = runnableCheck(c, opts);
    if (!r.runnable) {
      out.push(issue('PLAN_VERIFICATION_NOT_RUNNABLE',
        `## Verification ${c.label} can never run (${r.reason}): ${c.text}`,
        `Phase 7 runs every criterion and fells the run on one it cannot run. ${WAY_IN}`));
    }
  }
  const verified = new Set(criteria.filter(c => c.command).map(c => c.command.trim().replace(/\s+/g, ' ')));
  for (const s of sections) {
    const f = parseStepFields(s.body);
    if (f.has('Verify')) {
      const v = firstCommand(f.get('Verify').text);
      const r = runnableCheck(v, opts);
      if (!r.runnable) {
        out.push(issue('PLAN_STEP_VERIFY_NOT_RUNNABLE',
          `Step ${s.n}: \`Verify:\` does not open with one runnable test command (${r.reason})`,
          'The first backticked span of `Verify:` is the step\'s test command: proof of red and the done line run it, ' +
            `and nothing after it is run. ${WAY_IN}`));
      }
    }
    if (f.has('End-to-end') && BACKTICKED.test(f.get('End-to-end').text)) {
      const e = firstCommand(f.get('End-to-end').text);
      const r = runnableCheck(e, opts);
      if (!r.runnable) {
        out.push(issue('PLAN_END_TO_END_NOT_RUNNABLE',
          `Step ${s.n}: \`End-to-end:\` command can never run (${r.reason})`,
          `Run the product from a script: \`bash tests/e2e.sh\`. ${WAY_IN}`));
      } else if (!verified.has(e.command.trim().replace(/\s+/g, ' '))) {
        out.push(issue('PLAN_END_TO_END_NOT_VERIFIED',
          `Step ${s.n}: \`End-to-end:\` command \`${e.command}\` does not stand under ## Verification`,
          'Phase 7 runs `## Verification`; list the same command there, or the result is "not verified end to end".'));
      }
    }
  }
  return out;
}

/**
 * K10 and K11. `Verify:` is both the step's one test command and its proof of
 * red, so a step whose red test is a shell script could not also run the frozen
 * suite before its checkpoint; the plan wrote "run npm test before the
 * checkpoint" in prose that nothing enforced (voyage-lekerepo trial 2,
 * 2026-10-08). `Regression:` is that command, checked like `Verify:`.
 */
function checkRecovery(sections, opts) {
  const errors = [];
  const warnings = [];
  for (const s of sections) {
    const f = parseStepFields(s.body);
    if (f.has('Regression')) {
      const r = runnableCheck(firstCommand(f.get('Regression').text), opts);
      if (!r.runnable) {
        errors.push(issue('PLAN_STEP_REGRESSION_NOT_RUNNABLE',
          `Step ${s.n}: \`Regression:\` is not one runnable command (${r.reason})`,
          'Backtick ONE command the criteria runner allows, unpiped and without `&&` (`npm test`); ' +
            `the executor runs it green before the checkpoint. ${WAY_IN}`));
      }
    }
    if (f.has('On failure') && onFailureAction(f.get('On failure').text) === null) {
      warnings.push(issue('PLAN_STEP_ON_FAILURE_UNKNOWN',
        `Step ${s.n}: \`On failure:\` does not open with a recovery word the executor reads`,
        `Open with one of ${ON_FAILURE_ACTIONS.join(' | ')}, then the specifics (\`retry once, then escalate\`).`));
    }
    if (stepTestsOnly(s.body).testsOnly) {
      // The run-the-thing script alone is such a step too: `End-to-end:` rides on
      // the last step that builds behaviour (release review of 6.3.0, 2026-10-10).
      const e2e = f.has('End-to-end')
        ? ' `End-to-end:` rides on the last step that builds behaviour, with its script in that step\'s `Files:`, ' +
          'not on a step of its own.'
        : '';
      warnings.push(issue('PLAN_STEP_TESTS_ONLY',
        `Step ${s.n}: every file in \`Files:\` is a test; a step of only tests can never be proven red`,
        'The test for a behaviour belongs in the step that builds it: at this step\'s start point the behaviour ' +
          'already exists (an earlier step built it, or it was there before the plan), so its test passes there and ' +
          'red proof rejects it (`not-red-at-base`). Move the test into the step that builds the behaviour; a test ' +
          'meant to pass from the start (a golden or characterization test) is the `Regression:` of the step that ' +
          'changes the code, not a step of its own.' + e2e));
    }
  }
  return { errors, warnings };
}

/** Table body rows (skips the header and the |---| separator). */
function tableRows(text) {
  const rows = text.split('\n').filter(l => l.trim().startsWith('|'));
  return rows.slice(2).map(r => r.trim().replace(/^\||\|$/g, '').split('|').map(c => c.trim()));
}

function checkDesign(text, stepNumbers) {
  const out = [];
  const design = findSection(text, 'Design', 2);
  if (!design) {
    return [issue('PLAN_DESIGN_MISSING', 'Architectural plan without a `## Design` section',
      'Add the design part (templates/design-template.md), or plan under `ceremony: small|bounded` if the brief says so.')];
  }
  const subs = splitSections(design.text, 3);
  const titles = new Set(subs.filter(s => s.title).map(s => s.title.toLowerCase()));
  const missing = DESIGN_SUBSECTIONS.filter(t => !titles.has(t.toLowerCase()));
  if (missing.length > 0) {
    out.push(issue('PLAN_DESIGN_SECTION_MISSING', `## Design lacks: ${missing.join(', ')}`,
      `Required subsections: ${DESIGN_SUBSECTIONS.join(', ')}.`));
  }
  if (!findSection(text, 'Test Strategy', 2)) {
    out.push(issue('PLAN_DESIGN_SECTION_MISSING', 'Architectural plan lacks `## Test Strategy`',
      'The design part\'s test strategy lives in the plan\'s `## Test Strategy` section.'));
  }
  const sub = t => subs.find(s => s.title && s.title.toLowerCase() === t.toLowerCase());

  const map = sub('Module Map');
  if (map) {
    const blocks = fencedBlocks(map.text, 'mermaid');
    if (blocks.length === 0) {
      out.push(issue('PLAN_MODULE_MAP_EMPTY', '### Module Map has no ```mermaid block',
        'Draw the modules and their dependency direction as a mermaid graph.'));
    }
    // All blocks form one map: a cycle may run across two of them.
    const parsed = blocks.map(b => parseMermaid(b));
    const unparsed = parsed.flatMap(p => p.unparsed);
    if (unparsed.length > 0) {
      out.push(issue('PLAN_MODULE_MAP_UNPARSED', `Module map lines that look like links but give no edge: ${unparsed.join(' | ')}`,
        'Write links as `a --> b` (ids of letters, digits, _ . - /); the validator cannot check what it cannot read.'));
    }
    const cycle = findModuleCycle(parsed.flatMap(p => p.edges));
    if (cycle) {
      out.push(issue('PLAN_MODULE_CYCLE', `Module map has a dependency cycle: ${cycle.join(' -> ')}`,
        'Dependencies point one way. Break the cycle (extract the shared part, or invert one edge).'));
    }
  }

  const rules = sub('Architecture Rules');
  if (rules) {
    const rows = tableRows(rules.text);
    if (rows.length === 0) {
      out.push(issue('PLAN_ARCH_RULE_NO_ENFORCER', '### Architecture Rules lists no rule',
        'List at least one rule as a table row: | Rule | Enforced by | Proof |.'));
    }
    for (const row of rows) {
      const line = row.join(' | ');
      const checked = /checked by\s+`[^`]+`/i.test(line);
      const advisory = /\badvisory\b/i.test(line) && /\.claude\/rules\/[^\s`|]+\.md/.test(line);
      if (!checked && !advisory) {
        out.push(issue('PLAN_ARCH_RULE_NO_ENFORCER', `Rule without an enforcer: ${row[0] || line}`,
          'Each rule says "checked by `<command>`" or "advisory — `.claude/rules/<file>.md`".'));
        continue;
      }
      if (checked) {
        const proof = (row[2] ?? '').match(/Step (\d+)/);
        if (!proof || !stepNumbers.includes(Number(proof[1]))) {
          out.push(issue('PLAN_ARCH_RULE_NO_PROOF',
            `Checked rule without an existing proving step: ${row[0] || line}`,
            'Name the step (Proof column, "Step N") that shows the check pass, fail on a planted violation, and pass again.'));
        }
      }
    }
  }
  return out;
}

export function validatePlanContent(rawText, opts = {}) {
  const strict = opts.strict !== false;
  // CRLF plans (an editor on Windows) used to fail every manifest with
  // FM_SYNTAX: the YAML line parser does not expect a trailing \r.
  const text = typeof rawText === 'string' ? rawText.replace(/\r\n?/g, '\n') : rawText;
  const headRes = validatePlanHeadings(text, { strict });
  const errors = [...headRes.errors];
  const warnings = [...headRes.warnings];

  const steps = headRes.parsed?.steps || [];
  const sections = sliceSteps(text);
  const manRes = validateAllManifests(sections);
  errors.push(...manRes.errors);
  warnings.push(...manRes.warnings);

  if (steps.length > 0 && manRes.parsed.length !== steps.length) {
    errors.push(issue(
      'PLAN_MANIFEST_COUNT_MISMATCH',
      `Step count (${steps.length}) does not equal manifest count (${manRes.parsed.length})`,
    ));
  }

  const planVersion = extractPlanVersion(text);
  if (planVersion === null) {
    warnings.push(issue('PLAN_NO_VERSION', `No plan_version detected; current target is ${PLAN_VERSION_CURRENT}`));
  } else if (!PLAN_VERSIONS_SUPPORTED.includes(planVersion)) {
    warnings.push(issue('PLAN_VERSION_MISMATCH', `plan_version=${planVersion}, current target is ${PLAN_VERSION_CURRENT}`));
  }

  const recovery = checkRecovery(sections, { cwd: opts.cwd });
  if (strict) errors.push(...recovery.errors);
  else warnings.push(...recovery.errors);
  warnings.push(...recovery.warnings);

  // Steps a run appended (the review loop's remediation steps, /trekexecute
  // Phase 7.2) must be runnable steps whatever the plan's version: below 1.8
  // the step rules never ran, and a review step whose Verify is "re-run
  // /trekreview" validated and was executed (release review of 6.3.0,
  // 2026-10-10). From 1.8 the rules below already cover every step.
  const stepsFrom = Number.isInteger(opts.stepsFrom) && opts.stepsFrom > 0 ? opts.stepsFrom : null;
  if (stepsFrom !== null && !versionAtLeast(planVersion, '1.8')) {
    const appended = sections.filter((s) => Number(s.n) >= stepsFrom);
    const stepIssues = [
      ...checkSteps(appended).filter((i) => i.code !== 'PLAN_NO_END_TO_END'),
      ...checkRunnable('', appended, { cwd: opts.cwd }),
    ];
    for (const s of appended) {
      if (!parseStepFields(s.body).has('Verify')) {
        stepIssues.push(issue('PLAN_STEP_VERIFY_NOT_RUNNABLE', `Step ${s.n}: no \`Verify:\` field`,
          'An appended step needs one test command: proof of red and the done line run it.'));
      }
    }
    errors.push(...stepIssues);
  }

  const ceremony = extractPlanCeremony(text);
  if (versionAtLeast(planVersion, '1.8')) {
    const v18 = [...checkSteps(sections), ...checkRunnable(text, sections, { cwd: opts.cwd })];
    if (!ceremony.valid) {
      v18.push(issue('PLAN_CEREMONY_INVALID', `ceremony=${ceremony.raw}`,
        `Allowed: ${CEREMONY_VALUES.join(' | ')} (absent = architectural).`));
    } else if (ceremony.value === 'architectural') {
      v18.push(...checkDesign(text, steps.map(s => s.n)));
    }
    if (strict) errors.push(...v18);
    else warnings.push(...v18);
  }

  // v4.1 SC #20 — MANIFEST_PROFILE_DRIFT detection. Strict-mode only.
  // If plan frontmatter declares `profile: <name>` and any step manifest
  // declares `profile_used: <other>`, emit a warning (NOT an error) so
  // operators see drift but parsing remains forward-compat.
  if (strict) {
    const planFm = parseDocument(text).parsed?.frontmatter;
    const planProfile =
      planFm && typeof planFm.profile === 'string' ? planFm.profile : null;
    if (planProfile) {
      for (const m of manRes.parsed) {
        const stepProfile = m.manifest && m.manifest.profile_used;
        if (typeof stepProfile === 'string' && stepProfile !== planProfile) {
          warnings.push(issue(
            'MANIFEST_PROFILE_DRIFT',
            `step ${m.n}: profile_used = ${stepProfile}, plan profile = ${planProfile}`,
            'A step manifest declares a different profile than the plan frontmatter; ' +
              'verify whether this is intentional (manual override) or accidental drift.',
          ));
        }
      }
    }
  }

  return {
    valid: errors.length === 0,
    errors,
    warnings,
    parsed: { steps, manifests: manRes.parsed, planVersion, ceremony: ceremony.value },
  };
}

export function validatePlan(filePath, opts = {}) {
  if (!existsSync(filePath)) return fail(issue('PLAN_NOT_FOUND', `File not found: ${filePath}`));
  let text;
  try { text = readFileSync(filePath, 'utf-8'); }
  catch (e) { return fail(issue('PLAN_READ_ERROR', `Cannot read ${filePath}: ${e.message}`)); }
  const r = validatePlanContent(text, opts);
  return { ...r, parsed: { ...r.parsed, filePath } };
}

// realpath on both sides: `file://${argv[1]}` silently no-ops under a symlink
// or a path with a space (the CLI then exits 0 having checked nothing).
function isMain() {
  try {
    return realpathSync(process.argv[1]) === realpathSync(fileURLToPath(import.meta.url));
  } catch {
    return false;
  }
}

if (isMain()) {
  const args = process.argv.slice(2);
  const strict = !args.includes('--soft');
  const at = args.indexOf('--steps-from');
  const stepsFrom = at >= 0 ? Number(args[at + 1]) : null;
  const filePath = args.find((a, i) => !a.startsWith('--') && !(at >= 0 && i === at + 1));
  if (!filePath || (at >= 0 && !(Number.isInteger(stepsFrom) && stepsFrom > 0))) {
    process.stderr.write('Usage: plan-validator.mjs [--strict|--soft] [--json] [--steps-from <N>] <plan.md>\n');
    process.exit(2);
  }
  const r = validatePlan(filePath, { strict, stepsFrom });
  if (args.includes('--json')) {
    process.stdout.write(JSON.stringify({
      valid: r.valid,
      errors: r.errors,
      warnings: r.warnings,
      steps: r.parsed?.steps?.length ?? 0,
      planVersion: r.parsed?.planVersion ?? null,
      ceremony: r.parsed?.ceremony ?? null,
    }, null, 2) + '\n');
  } else {
    process.stdout.write(`plan-validator: ${r.valid ? 'READY' : 'FAIL'} ${filePath} (${r.parsed?.steps?.length ?? 0} steps)\n`);
    for (const e of r.errors) process.stderr.write(`  ERROR  [${e.code}] ${e.message}\n`);
    for (const w of r.warnings) process.stderr.write(`  WARN   [${w.code}] ${w.message}\n`);
  }
  process.exit(r.valid ? 0 : 1);
}