principles-disciple 1.81.0 → 1.83.0

package/templates/langs/zh/skills/ai-sprint-orchestration/references/workflow-v1-acceptance-checklist.md

@@ -1,58 +0,0 @@
-# Workflow v1 Acceptance Checklist
-
-## Purpose
-
-Use this checklist to validate that the packaged workflow is readable, executable, and safe to hand off to another agent.
-
-## Acceptance gates
-
-- [ ] Baseline tests are green
-- [ ] `workflow-validation-minimal` completes and writes package-local artifacts
-- [ ] `workflow-validation-minimal-verify` completes and validates the previous run
-- [ ] Every failure is classified into the approved four-category taxonomy
-
-## Commands
-
-```powershell
-node scripts/run.mjs --self-check
-node scripts/run.mjs --help
-node scripts/run.mjs --task workflow-validation-minimal
-node scripts/run.mjs --task workflow-validation-minimal-verify
-```
-
-## Run result record
-
-| Field | Value |
-|------|------|
-| runId | |
-| outcome | |
-| outputQuality | |
-| validation | |
-| nextRunRecommendation | |
-| failure classification | |
-
-## Required artifact checks
-
-- `decision.md` contains `outputQuality`
-- `decision.md` contains `qualityReasons`
-- `decision.md` contains readable validation status
-- `scorecard.json` contains `outputQuality`
-- `scorecard.json` contains `qualityReasons`
-- `scorecard.json` contains `validation`
-- `scorecard.json` contains `nextRunRecommendation`
-
-## Failure classification
-
-- `workflow bug`
-- `agent behavior issue`
-- `environment issue`
-- `sample-spec issue`
-
-## Stop conditions
-
-If the run exposes a sample-side or product-side issue:
-
-- classify it
-- record evidence
-- stop the run review
-- do not continue into product closure work

package/templates/langs/zh/skills/ai-sprint-orchestration/references/workflow-v1.4-work-unit-handoff.md

@@ -1,190 +0,0 @@
-# Workflow v1.4 Work-Unit Upgrade Handoff
-
-## Goal
-
-Use the packaged `ai-sprint-orchestration` skill to implement the next architecture upgrade of the workflow itself.
-
-This upgrade is **not** about product-side closure. It is about making the orchestrator better at long-running, high-value coding tasks by introducing a finer-grained execution model.
-
-The target direction is:
-
-- move from `stage -> producer/reviewer`
-- toward `stage -> work units -> producer/reviewer/decision`
-
-The key design goal is:
-
-- smaller execution units
-- explicit checkpointing
-- forced context reload between units
-- less long-context drift
-
-## Current confirmed state
-
-The packaged skill is already usable as an internal tool:
-
-- package path:
-  - `D:/Code/principles/packages/openclaw-plugin/templates/langs/zh/skills/ai-sprint-orchestration`
-- installed path:
-  - `C:/Users/Administrator/.agents/skills/ai-sprint-orchestration`
-- self-check works
-- package-local validation runs work
-- failure classification is persisted to:
-  - `latest-summary.md`
-  - `scorecard.json`
-- complex task templates and minimum task contract already exist
-
-Remaining instability is mostly:
-
-- `agent behavior issue`
-
-It is **not** primarily:
-
-- pathing
-- packaging
-- runtime-root layout
-- missing acceptance artifacts
-
-## Why v1.4 is needed
-
-The current workflow already externalizes state and can resume from artifacts, but its granularity is still too coarse for very complex tasks.
-
-Today it mainly refreshes context at:
-
-- stage level
-- role level
-- round level
-
-That is useful, but not yet enough for difficult long-running work where one producer pass can still become too large, too noisy, or too drift-prone.
-
-The next upgrade should add a smaller unit below stage:
-
-- `workUnitId`
-- `workUnitGoal`
-- `allowedFiles`
-- `unitChecks`
-- `unitDeliverables`
-- `unitSummary`
-- `carryForwardSummary`
-
-## Scope
-
-### In scope
-
-- `packages/openclaw-plugin/templates/langs/zh/skills/ai-sprint-orchestration`
-- `packages/openclaw-plugin/templates/langs/en/skills/ai-sprint-orchestration`
-- workflow docs, specs, prompts, state shape, and summary artifacts
-- package-local validation and workflow-only tests
-
-### Out of scope
-
-- `packages/openclaw-plugin`
-- `D:/Code/openclaw`
-- product-side/sample-side closure
-- dashboard
-- stageGraph
-- self-optimizing sprint
-- parallel orchestrator expansion
-
-If validation or implementation exposes product-side or sample-side issues, classify them and stop. Do not drift back into product closure work.
-
-## Architectural hypothesis
-
-The most valuable next step is **not** a full rewrite.
-
-The most valuable next step is to implement a high-value foundation slice:
-
-1. define the work-unit contract
-2. define how a stage declares work units
-3. add package-local templates and docs for work-unit-aware tasks
-4. add checkpoint/carry-forward artifacts that are short enough to be reused safely
-5. update prompts so each unit starts from minimal context, not long historical text
-
-This means the first v1.4 sprint should prioritize:
-
-- work-unit interfaces
-- unit-level artifacts
-- unit-level carry-forward
-- validation of the new context-reload behavior
-
-It should **not** try to finish every future v1.4 feature in one pass.
-
-## Recommended implementation order
-
-### Phase 1: foundation interfaces
-
-Define and document the unit contract in specs and references:
-
-- `workUnitId`
-- `workUnitGoal`
-- `allowedFiles`
-- `unitChecks`
-- `unitDeliverables`
-- `unitSummary`
-- `carryForwardSummary`
-
-Add minimal schema/contract validation for these fields where appropriate.
-
-### Phase 2: unit-scoped prompt inputs
-
-Update producer/reviewer prompt construction so a unit run receives:
-
-- current unit goal
-- allowed files
-- expected checks
-- expected deliverables
-- prior carry-forward summary
-
-Prefer short checkpoint summaries over replaying long decision history.
-
-### Phase 3: checkpoint and continuation tightening
-
-Standardize the artifact that one unit leaves behind for the next:
-
-- accomplished
-- blockers
-- next focus
-- verified files
-
-Make continuation/revise paths prefer this short artifact.
-
-### Phase 4: validation
-
-Add or update validation specs and tests so the workflow can prove:
-
-- missing task contract is rejected
-- work-unit metadata is loaded correctly
-- continuation reads compact carry-forward context
-- failures still classify cleanly
-
-## Success criteria
-
-The first v1.4 slice is successful if all of these are true:
-
-1. The workflow can express work-unit metadata in package-local specs or templates.
-2. Prompts clearly scope a run to a smaller work unit, not a broad stage-only brief.
-3. Continuation uses compact carry-forward summaries by default.
-4. Baseline tests pass.
-5. Package-local self-check still passes.
-6. Validation runs either pass, or fail with explicit classification that is not a workflow-plumbing ambiguity.
-
-## Guardrails
-
-- Do not rewrite the orchestrator from scratch.
-- Do not introduce a second orchestration system.
-- Do not move back into product/sample closure.
-- Do not increase scope to unrelated features.
-- Prefer one workflow-only architectural improvement per iteration.
-- Preserve the packaged skill as the primary operator entry point.
-
-## Suggested starting point for the next thread
-
-In the next thread, the first sprint should focus on:
-
-- implementing the work-unit contract and carry-forward foundation
-- not the entire future v1.4 roadmap
-
-Treat this as a bounded architecture slice:
-
-- high value
-- low drift
-- directly useful for future complex coding tasks

package/templates/langs/zh/skills/ai-sprint-orchestration/runtime/.gitignore

@@ -1,2 +0,0 @@
-*
-!.gitignore

package/templates/langs/zh/skills/ai-sprint-orchestration/scripts/lib/archive.mjs

@@ -1,310 +0,0 @@
-import fs from 'fs';
-import path from 'path';
-import { fileURLToPath } from 'url';
-import { ensureDir, readJson, writeJson, writeText, fileExists } from './state-store.mjs';
-
-const __dirname = path.dirname(fileURLToPath(import.meta.url));
-const packageRoot = path.resolve(__dirname, '..', '..');
-
-function resolveRoots() {
-  const runtimeRoot = process.env.AI_SPRINT_RUNTIME_ROOT
-    ? path.resolve(process.env.AI_SPRINT_RUNTIME_ROOT)
-    : path.join(packageRoot, 'runtime');
-  return {
-    runtimeRoot,
-    sprintRoot: path.join(runtimeRoot, 'runs'),
-    archiveRoot: path.join(runtimeRoot, 'archive'),
-  };
-}
-
-function nowIso() {
-  return new Date().toISOString();
-}
-
-function shouldArchiveEntry(srcPath, runDir) {
-  const relative = path.relative(runDir, srcPath);
-  if (!relative || relative === '') return true;
-
-  const normalized = relative.split(path.sep).join('/');
-  const base = path.basename(srcPath);
-
-  if (base === 'orchestrator.lock') return false;
-  if (base.startsWith('.ai-sprint-prompt-')) return false;
-  if (normalized.includes('/worktrees/')) return false;
-  if (normalized.startsWith('worktrees/')) return false;
-  if (normalized.includes('/runtime/')) return false;
-  if (normalized.startsWith('runtime/')) return false;
-
-  return true;
-}
-
-/**
- * Archive a completed/halted sprint run by run ID.
- * CLI entry point for --archive <run-id>.
- */
-export function archiveRunById(runId) {
-  const { sprintRoot } = resolveRoots();
-  const runDir = path.join(sprintRoot, runId);
-  const sprintFile = path.join(runDir, 'sprint.json');
-  if (!fileExists(sprintFile)) {
-    throw new Error(`Run not found: ${runId}`);
-  }
-  return archiveRun(runDir, runId);
-}
-
-/**
- * Core archive function. Copies all artifacts, captures git info, generates summary.
- * @param {string} runDir - Path to the sprint run directory
- * @param {string} runId - The run ID
- * @returns {string} Path to the archive directory
- */
-export function archiveRun(runDir, runId) {
-  const { archiveRoot } = resolveRoots();
-  const destDir = path.join(archiveRoot, runId);
-
-  // Idempotency: check if already successfully archived
-  const metaPath = path.join(destDir, 'archive-meta.json');
-  if (fileExists(metaPath)) {
-    const meta = readJson(metaPath);
-    if (meta.status === 'completed') {
-      throw new Error(`Already archived: ${destDir}`);
-    }
-    // Partial/failed archive — clean up and re-run
-    fs.rmSync(destDir, { recursive: true, force: true });
-  }
-
-  // Read sprint state
-  const state = readJson(path.join(runDir, 'sprint.json'));
-
-  // Warn if archiving a live run
-  if (state.status === 'running' || state.status === 'paused') {
-    console.error(`Warning: run ${runId} has status '${state.status}'. Archiving a live run may produce incomplete artifacts.`);
-  }
-
-  // Step 1: Copy all run artifacts recursively
-  ensureDir(destDir);
-  try {
-    fs.cpSync(runDir, destDir, {
-      recursive: true,
-      filter: (src) => shouldArchiveEntry(src, runDir),
-    });
-  } catch {
-    // Fallback for Node < 16.7: manual recursive copy
-    copyDirRecursive(runDir, destDir, runDir);
-  }
-
-  // Step 2: Capture git info
-  captureGitInfo(destDir, state);
-
-  // Step 3: Generate summary
-  generateSummary(destDir, state);
-
-  // Step 4: Write archive metadata (last — marks archive as complete)
-  writeJson(metaPath, {
-    runId,
-    archivedAt: nowIso(),
-    status: 'completed',
-    sourceStatus: state.status,
-  });
-
-  return destDir;
-}
-
-/**
- * Capture git information into the archive's git/ directory.
- * Each command is independent — one failure does not block the rest.
- */
-function captureGitInfo(destDir, state) {
-  const gitDir = path.join(destDir, 'git');
-  ensureDir(gitDir);
-  const latestGitStatus = findLatestStageGitStatus(destDir);
-
-  if (!latestGitStatus) {
-    writeText(path.join(gitDir, 'branch.txt'), '# No stage git-status.json found in archive\n');
-    writeText(path.join(gitDir, 'status.txt'), '# No stage git-status.json found in archive\n');
-    writeText(path.join(gitDir, 'modified-files.txt'), '# No stage git-status.json found in archive\n');
-    writeText(path.join(gitDir, 'diff.patch'), '# Diff omitted from archive; inspect archived stage artifacts instead.\n');
-    writeText(path.join(gitDir, 'log.txt'), '# Git log omitted from archive; inspect archived stage artifacts instead.\n');
-    return;
-  }
-
-  const dirtyLines = Array.isArray(latestGitStatus.dirtyFiles) ? latestGitStatus.dirtyFiles : [];
-  const modifiedFiles = dirtyLines
-    .map((line) => String(line).replace(/^[ MARCUD?!]+/, '').trim())
-    .filter(Boolean);
-
-  writeText(path.join(gitDir, 'branch.txt'), `${latestGitStatus.branch || 'unknown'}\n`);
-  writeText(path.join(gitDir, 'status.txt'), `${dirtyLines.length ? dirtyLines.join('\n') : '# clean'}\n`);
-  writeText(path.join(gitDir, 'modified-files.txt'), `${modifiedFiles.length ? modifiedFiles.join('\n') : '# none'}\n`);
-  writeText(path.join(gitDir, 'diff.patch'), '# Diff omitted from archive; inspect archived stage artifacts and git-status.json.\n');
-  writeText(path.join(gitDir, 'log.txt'), [
-    `headSha: ${latestGitStatus.headSha || 'unknown'}`,
-    `baseBranch: ${latestGitStatus.baseBranch || 'unknown'}`,
-    `remoteBranch: ${latestGitStatus.remoteBranch || 'unknown'}`,
-    `capturedAt: ${state.updatedAt || nowIso()}`,
-  ].join('\n') + '\n');
-}
-
-function findLatestStageGitStatus(destDir) {
-  const stagesDir = path.join(destDir, 'stages');
-  if (!fileExists(stagesDir)) return null;
-
-  const stageDirs = fs.readdirSync(stagesDir).sort().reverse();
-  for (const dir of stageDirs) {
-    const gitStatusPath = path.join(stagesDir, dir, 'git-status.json');
-    if (fileExists(gitStatusPath)) {
-      return readJson(gitStatusPath);
-    }
-  }
-  return null;
-}
-
-/**
- * Generate a human-readable archive summary.
- */
-function generateSummary(destDir, state) {
-  const lines = [];
-
-  // Identity
-  lines.push(`# Sprint Archive: ${state.title || state.taskId}`);
-  lines.push('');
-  lines.push('## Identity');
-  lines.push(`- Run ID: ${state.runId}`);
-  lines.push(`- Task: ${state.taskId}`);
-  lines.push(`- Status: ${state.status}`);
-  lines.push(`- Archived at: ${nowIso()}`);
-  lines.push('');
-
-  // Timeline
-  const created = state.createdAt || '';
-  const updated = state.updatedAt || '';
-  const elapsedMs = created ? (Date.parse(updated) || Date.now()) - Date.parse(created) : 0;
-  const elapsedMin = (elapsedMs / 60_000).toFixed(1);
-  lines.push('## Timeline');
-  lines.push(`- Created: ${created}`);
-  lines.push(`- Updated: ${updated}`);
-  lines.push(`- Total wall time: ${elapsedMin} minutes`);
-  lines.push(`- Final stage: ${state.currentStage} (index ${state.currentStageIndex})`);
-  lines.push(`- Final round: ${state.currentRound}`);
-  lines.push('');
-
-  // Stage progress table from scorecards
-  lines.push('## Stage Progress');
-  const stagesDir = path.join(destDir, 'stages');
-  const stageEntries = [];
-  if (fileExists(stagesDir)) {
-    const dirs = fs.readdirSync(stagesDir).sort();
-    for (const dir of dirs) {
-      const scorecardPath = path.join(stagesDir, dir, 'scorecard.json');
-      if (fileExists(scorecardPath)) {
-        const sc = readJson(scorecardPath);
-        stageEntries.push({
-          dir,
-          outcome: sc.outcome || '?',
-          round: sc.round || '?',
-          approvals: sc.approvalCount ?? '?',
-          blockers: sc.blockerCount ?? 0,
-          reviewerA: sc.reviewerAVerdict || '?',
-          reviewerB: sc.reviewerBVerdict || '?',
-        });
-      }
-    }
-  }
-
-  if (stageEntries.length > 0) {
-    lines.push('| Stage | Outcome | Round | Approvals | Blockers | Reviewer A | Reviewer B |');
-    lines.push('|-------|---------|-------|-----------|----------|-----------|-----------|');
-    for (const s of stageEntries) {
-      lines.push(`| ${s.dir} | ${s.outcome} | ${s.round} | ${s.approvals}/2 | ${s.blockers} | ${s.reviewerA} | ${s.reviewerB} |`);
-    }
-  } else {
-    lines.push('No stage scorecards found.');
-  }
-  lines.push('');
-
-  // Git context
-  lines.push('## Git Context');
-  const gitDir = path.join(destDir, 'git');
-  if (fileExists(path.join(gitDir, 'branch.txt'))) {
-    lines.push(`- Branch: ${fs.readFileSync(path.join(gitDir, 'branch.txt'), 'utf8').trim()}`);
-  }
-  if (fileExists(path.join(gitDir, 'log.txt'))) {
-    const log = fs.readFileSync(path.join(gitDir, 'log.txt'), 'utf8').trim();
-    if (log && !log.startsWith('#')) {
-      lines.push('');
-      lines.push('### Commits');
-      for (const line of log.split('\n').filter(Boolean)) {
-        lines.push(`- ${line}`);
-      }
-    }
-  }
-  if (fileExists(path.join(gitDir, 'modified-files.txt'))) {
-    const files = fs.readFileSync(path.join(gitDir, 'modified-files.txt'), 'utf8').trim();
-    if (files && !files.startsWith('#')) {
-      lines.push('');
-      lines.push('### Modified Files');
-      for (const f of files.split('\n').filter(Boolean)) {
-        lines.push(`- ${f}`);
-      }
-    }
-  }
-  lines.push('');
-
-  // Halt reason
-  if (state.haltReason) {
-    lines.push('## Halt Reason');
-    lines.push(`- Type: ${state.haltReason.type}`);
-    lines.push(`- Details: ${state.haltReason.details}`);
-    if (state.haltReason.blockers?.length) {
-      lines.push('### Blockers');
-      for (const b of state.haltReason.blockers) {
-        lines.push(`- ${b}`);
-      }
-    }
-    lines.push('');
-  }
-
-  // Open risks from verify producer
-  const verifyDirs = stageEntries.filter((s) => s.dir.includes('verify'));
-  if (verifyDirs.length > 0) {
-    const verifyDir = path.join(stagesDir, verifyDirs[verifyDirs.length - 1].dir);
-    const producerPath = path.join(verifyDir, 'producer.md');
-    if (fileExists(producerPath)) {
-      const producerText = fs.readFileSync(producerPath, 'utf8');
-      const risksSection = extractSection(producerText, 'OPEN_RISKS');
-      if (risksSection) {
-        lines.push('## Open Risks (from verify producer)');
-        lines.push(risksSection);
-        lines.push('');
-      }
-    }
-  }
-
-  writeText(path.join(destDir, 'archive-summary.md'), lines.join('\n') + '\n');
-}
-
-/**
- * Extract a markdown section body by heading name.
- */
-function extractSection(text, heading) {
-  const pattern = new RegExp(`^##\\s+${heading}\\s*\\n([\\s\\S]*?)(?=^##\\s|$(?!\\n))`, 'm');
-  const match = text.match(pattern);
-  return match ? match[1].trim() : null;
-}
-
-/**
- * Fallback recursive directory copy for Node < 16.7.
- */
-function copyDirRecursive(src, dest, rootDir = src) {
-  ensureDir(dest);
-  for (const entry of fs.readdirSync(src, { withFileTypes: true })) {
-    const srcPath = path.join(src, entry.name);
-    const destPath = path.join(dest, entry.name);
-    if (!shouldArchiveEntry(srcPath, rootDir)) continue;
-    if (entry.isDirectory()) {
-      copyDirRecursive(srcPath, destPath, rootDir);
-    } else {
-      fs.copyFileSync(srcPath, destPath);
-    }
-  }
-}