karajan-code 4.70.3 → 5.0.0

package/src/utils/status-dashboard.js

@@ -1,254 +0,0 @@
-/**
- * Status dashboard formatter.
- * Builds a structured terminal dashboard from session data and run-log lines.
- */
-
-import { HU_STATUS } from "../hu/store.js";
-
-/**
- * Status label mapping with visual indicators for terminal output.
- * @type {Record<string, string>}
- */
-const STATUS_LABELS = {
-  [HU_STATUS.PENDING]:       "pending",
-  [HU_STATUS.CODING]:        "coding",
-  [HU_STATUS.REVIEWING]:     "reviewing",
-  [HU_STATUS.DONE]:          "done",
-  [HU_STATUS.FAILED]:        "failed",
-  [HU_STATUS.BLOCKED]:       "blocked",
-  [HU_STATUS.CERTIFIED]:     "certified",
-  [HU_STATUS.NEEDS_CONTEXT]: "needs_ctx",
-};
-
-/**
- * Format a millisecond duration into a human-readable string (e.g. "3m 42s").
- * @param {number} ms
- * @returns {string}
- */
-export function formatDuration(ms) {
-  if (!ms || ms < 0) return "0s";
-  const totalSec = Math.floor(ms / 1000);
-  const hours = Math.floor(totalSec / 3600);
-  const minutes = Math.floor((totalSec % 3600) / 60);
-  const seconds = totalSec % 60;
-
-  const parts = [];
-  if (hours > 0) parts.push(`${hours}h`);
-  if (minutes > 0) parts.push(`${minutes}m`);
-  if (seconds > 0 || parts.length === 0) parts.push(`${seconds}s`);
-  return parts.join(" ");
-}
-
-/**
- * Parse the current pipeline stage from log lines.
- * @param {string[]} lines - Raw log lines.
- * @returns {{ currentStage: string|null, lastEvent: string|null }}
- */
-export function parseStageFromLog(lines) {
-  let currentStage = null;
-  let lastEvent = null;
-
-  for (const line of lines) {
-    const stageStart = /\[(\w+):start\]/.exec(line);
-    if (stageStart) currentStage = stageStart[1];
-
-    const stageDone = /\[(\w+):done\]|\[(\w+)\] finished/.exec(line);
-    if (stageDone) {
-      const doneName = stageDone[1] || stageDone[2];
-      if (doneName === currentStage) currentStage = "idle";
-    }
-
-    if (line.includes("[standby]") || line.includes("standby")) {
-      currentStage = "standby";
-    }
-
-    if (line.trim()) lastEvent = line.trim();
-  }
-
-  return { currentStage, lastEvent };
-}
-
-/**
- * Format a single HU line for the dashboard.
- * @param {object} story - HU story object.
- * @param {boolean} isCurrent - Whether this is the currently active HU.
- * @returns {string}
- */
-function formatHuLine(story, isCurrent) {
-  const id = (story.id || "???").padEnd(12);
-  const rawStatus = story.status || "pending";
-  const label = STATUS_LABELS[rawStatus] || rawStatus;
-  const statusStr = `[${label}]`.padEnd(13);
-  const title = story.title || story.original?.text || "(untitled)";
-  const truncTitle = title.length > 40 ? title.slice(0, 37) + "..." : title;
-
-  let timing = "";
-  if (story.duration_ms) {
-    timing = formatDuration(story.duration_ms);
-  }
-
-  const blockedInfo = story.blocked_by?.length
-    ? ` (needs ${story.blocked_by.join(", ")})`
-    : "";
-
-  const marker = isCurrent ? "  <-- current" : "";
-  const timingStr = timing ? `  ${timing}` : "";
-
-  return `  ${id} ${statusStr} ${truncTitle}${blockedInfo}${timingStr}${marker}`;
-}
-
-/**
- * Build a terminal dashboard string from session data and log lines.
- *
- * @param {object|null} sessionData - Session JSON (from session-store) or null.
- * @param {string[]} logLines - Raw run-log lines.
- * @param {object} [options] - Optional overrides.
- * @param {object[]|null} [options.stories] - HU stories array (from hu batch).
- * @returns {string}
- */
-export function buildDashboard(sessionData, logLines, options = {}) {
-  if (!sessionData) {
-    return "No active pipeline";
-  }
-
-  const lines = [];
-  const config = sessionData.config_snapshot || {};
-  const maxIter = config.max_iterations ?? "?";
-  const status = sessionData.status || "unknown";
-
-  // Parse log for current stage info
-  const logInfo = parseStageFromLog(logLines || []);
-
-  // Determine iteration from session or log
-  const iteration = sessionData.reviewer_retry_count != null
-    ? sessionData.reviewer_retry_count + 1
-    : null;
-  const iterFromLog = parseIterationFromLog(logLines || []);
-  const displayIter = iterFromLog || iteration || "?";
-
-  // Pipeline header
-  const statusUpper = status.toUpperCase();
-  lines.push(`Pipeline: ${statusUpper} (iteration ${displayIter}/${maxIter})`);
-
-  // Duration
-  const duration = computeDuration(sessionData);
-  lines.push(`Duration: ${formatDuration(duration)}`);
-
-  // Agent info
-  const coderProvider = config.coder?.provider || config.roles?.coder?.provider || "?";
-  const reviewerProvider = config.reviewer?.provider || config.roles?.reviewer?.provider || "?";
-  lines.push(`Agent: ${coderProvider} (coder) / ${reviewerProvider} (reviewer)`);
-
-  // HU stories section
-  const stories = options.stories || [];
-  if (stories.length > 0) {
-    lines.push("");
-    lines.push("HUs:");
-
-    // Determine which story is currently being worked on
-    const currentStory = stories.find(s =>
-      s.status === HU_STATUS.CODING || s.status === HU_STATUS.REVIEWING
-    );
-
-    for (const story of stories) {
-      const isCurrent = currentStory && story.id === currentStory.id;
-      lines.push(formatHuLine(story, isCurrent));
-    }
-  }
-
-  // Current stage and last event
-  lines.push("");
-  const stage = logInfo.currentStage || "idle";
-  const iterSuffix = displayIter !== "?" ? ` (iteration ${displayIter})` : "";
-  lines.push(`Current stage: ${stage}${iterSuffix}`);
-
-  if (logInfo.lastEvent) {
-    // Truncate last event for display
-    const maxLen = 120;
-    const event = logInfo.lastEvent.length > maxLen
-      ? logInfo.lastEvent.slice(0, maxLen - 3) + "..."
-      : logInfo.lastEvent;
-    lines.push(`Last event: ${event}`);
-  }
-
-  // Errors
-  if (sessionData.deferred_issues?.length) {
-    lines.push("");
-    lines.push(`Deferred issues: ${sessionData.deferred_issues.length}`);
-  }
-
-  return lines.join("\n");
-}
-
-/**
- * Build a structured JSON response for MCP consumers.
- *
- * @param {object|null} sessionData - Session JSON or null.
- * @param {string[]} logLines - Raw run-log lines.
- * @param {object} [options] - Optional overrides.
- * @param {object[]|null} [options.stories] - HU stories array.
- * @returns {object}
- */
-export function buildDashboardJson(sessionData, logLines, options = {}) {
-  if (!sessionData) {
-    return { ok: false, message: "No active pipeline" };
-  }
-
-  const config = sessionData.config_snapshot || {};
-  const logInfo = parseStageFromLog(logLines || []);
-  const iterFromLog = parseIterationFromLog(logLines || []);
-  const iteration = iterFromLog
-    || (sessionData.reviewer_retry_count != null ? sessionData.reviewer_retry_count + 1 : null);
-
-  const stories = (options.stories || []).map(s => ({
-    id: s.id,
-    status: s.status || "pending",
-    title: s.title || s.original?.text || null,
-    blocked_by: s.blocked_by || [],
-    duration_ms: s.duration_ms || null,
-  }));
-
-  return {
-    ok: true,
-    pipeline: {
-      status: sessionData.status || "unknown",
-      iteration: iteration || null,
-      maxIterations: config.max_iterations ?? null,
-      duration_ms: computeDuration(sessionData),
-      coder: config.coder?.provider || config.roles?.coder?.provider || null,
-      reviewer: config.reviewer?.provider || config.roles?.reviewer?.provider || null,
-    },
-    hus: stories,
-    currentStage: logInfo.currentStage || "idle",
-    lastEvent: logInfo.lastEvent || null,
-  };
-}
-
-/**
- * Parse iteration number from log lines.
- * @param {string[]} lines
- * @returns {number|null}
- */
-function parseIterationFromLog(lines) {
-  let iteration = null;
-  for (const line of lines) {
-    const m = /[Ii]teration\s+(\d+)/.exec(line);
-    if (m) iteration = Number.parseInt(m[1], 10);
-  }
-  return iteration;
-}
-
-/**
- * Compute pipeline duration in milliseconds from session timestamps.
- * @param {object} sessionData
- * @returns {number}
- */
-function computeDuration(sessionData) {
-  const created = sessionData.created_at;
-  if (!created) return 0;
-  const start = new Date(created).getTime();
-  const now = sessionData.updated_at
-    ? new Date(sessionData.updated_at).getTime()
-    : Date.now();
-  return Math.max(0, now - start);
-}

package/src/utils/tool-call-extractor.js

@@ -1,56 +0,0 @@
-// KJC-TSK-0375 PR2 — Extract structured tool calls from an agent transcript.
-//
-// The ToolCorrectnessJudgeRole (PR1) consumes a list of `{tool, args}`
-// objects. The coder writes a raw transcript (provider-dependent format) to
-// disk per iteration. This module bridges the two: parse the transcript,
-// return the ordered list of tool invocations.
-//
-// Provider-agnostic interface; Claude stream-json is the only format
-// implemented today. New providers add a branch below — keep parsers
-// pure (no I/O), so the role tests can mock transcripts trivially.
-
-const PROVIDER_PARSERS = {
-  claude: parseClaudeStreamJson,
-};
-
-/**
- * Extract tool calls from a transcript string.
- *
- * @param {object} args
- * @param {string} args.transcript - raw agent output (stdout/stderr buffer).
- * @param {string} [args.provider] - "claude" (default). Unknown returns [].
- * @returns {Array<{tool:string, args:object}>}
- */
-export function extractToolCalls({ transcript, provider = "claude" } = {}) {
-  if (typeof transcript !== "string" || transcript.length === 0) return [];
-  const parser = PROVIDER_PARSERS[provider];
-  if (!parser) return [];
-  return parser(transcript);
-}
-
-// Claude emits one JSON event per line. We care about `assistant` events
-// whose `message.content` is an array containing `tool_use` blocks with
-// `{type:"tool_use", name, input}`. Lines that fail to parse, that are
-// not `assistant`, or whose content is not an array are silently skipped:
-// the transcript also carries `system`, `result`, `rate_limit_event`, etc.
-function parseClaudeStreamJson(transcript) {
-  const out = [];
-  for (const raw of transcript.split("\n")) {
-    const line = raw.trim();
-    if (!line || line[0] !== "{") continue;
-    let obj;
-    try { obj = JSON.parse(line); } catch { continue; }
-    if (obj?.type !== "assistant") continue;
-    const content = obj?.message?.content;
-    if (!Array.isArray(content)) continue;
-    for (const block of content) {
-      if (block?.type === "tool_use" && typeof block.name === "string") {
-        out.push({
-          tool: block.name,
-          args: block.input && typeof block.input === "object" ? block.input : {},
-        });
-      }
-    }
-  }
-  return out;
-}

package/src/webperf/devtools-detect.js

@@ -1,112 +0,0 @@
-// Chrome DevTools MCP detection and WebPerf skill auto-installation.
-
-import fs from "node:fs/promises";
-import path from "node:path";
-import os from "node:os";
-import { isOpenSkillsAvailable, listSkills } from "../skills/openskills-client.js";
-
-/** Skill names relevant for WebPerf analysis. Already installed globally via openskills. */
-export const WEBPERF_SKILLS = [
-  "optimize",
-  "audit",
-  "polish"
-];
-
-/**
- * Check whether Chrome DevTools MCP is configured and available.
- * Checks: (1) config flag, (2) Claude MCP config (~/.claude.json), (3) VS Code MCP settings.
- * @param {object} config - merged Karajan config
- * @returns {boolean}
- */
-export function isDevToolsMcpAvailable(config) {
-  // Explicit config flag takes priority
-  if (config?.webperf?.devtools_mcp === true) return true;
-  if (config?.webperf?.devtools_mcp === false) return false;
-
-  // Check cached detection result
-  if (_devToolsDetected !== null) return _devToolsDetected;
-
-  // Not yet detected - return false (async detection runs at init)
-  return false;
-}
-
-let _devToolsDetected = null;
-
-/**
- * Auto-detect Chrome DevTools MCP from system configuration.
- * Checks ~/.claude.json and VS Code MCP settings.
- * Call once at pipeline init.
- */
-export async function detectDevToolsMcp(logger) {
-  // Check Claude MCP config
-  const claudeConfigPath = path.join(os.homedir(), ".claude.json");
-  try {
-    const raw = await fs.readFile(claudeConfigPath, "utf8");
-    const parsed = JSON.parse(raw);
-    const mcpServers = parsed?.mcpServers || parsed?.mcp_servers || {};
-    if (mcpServers["chrome-devtools"] || mcpServers["chrome_devtools"]) {
-      _devToolsDetected = true;
-      logger?.info?.("Chrome DevTools MCP detected in ~/.claude.json");
-      return true;
-    }
-  } catch { /* file not found or parse error */ }
-
-  // Check VS Code MCP settings
-  const vscodePaths = [
-    path.join(os.homedir(), ".vscode", "settings.json"),
-    path.join(process.cwd(), ".vscode", "settings.json")
-  ];
-  for (const vscPath of vscodePaths) {
-    try {
-      const raw = await fs.readFile(vscPath, "utf8");
-      if (raw.includes("chrome-devtools") || raw.includes("chrome_devtools")) {
-        _devToolsDetected = true;
-        logger?.info?.(`Chrome DevTools MCP detected in ${vscPath}`);
-        return true;
-      }
-    } catch { /* not found */ }
-  }
-
-  _devToolsDetected = false;
-  return false;
-}
-
-/**
- * Ensure WebPerf skills are installed via OpenSkills.
- * Installs any missing skills from WEBPERF_SKILLS.
- * @param {string} projectDir - absolute path to the project root
- * @param {object} [logger] - optional logger ({ info, warn })
- * @returns {Promise<{ installed: string[], alreadyInstalled: string[], skipped: string[] }>}
- */
-export async function ensureWebPerfSkills(projectDir, logger) {
-  const result = { installed: [], alreadyInstalled: [], skipped: [] };
-
-  const available = await isOpenSkillsAvailable();
-  if (!available) {
-    logger?.debug?.("OpenSkills CLI not available - WebPerf skills unavailable");
-    result.skipped = [...WEBPERF_SKILLS];
-    return result;
-  }
-
-  // List skills (includes global installs)
-  const listResult = await listSkills({ projectDir });
-  const installedNames = new Set(
-    (listResult.ok ? listResult.skills : []).map(s => s.name)
-  );
-
-  for (const skill of WEBPERF_SKILLS) {
-    if (installedNames.has(skill)) {
-      result.alreadyInstalled.push(skill);
-    } else {
-      // Skills should be installed globally via `openskills install owner/repo`
-      // Don't attempt auto-install with bare names — just note they're missing
-      result.skipped.push(skill);
-    }
-  }
-
-  if (result.alreadyInstalled.length > 0) {
-    logger?.info?.(`WebPerf skills available: ${result.alreadyInstalled.join(", ")}`);
-  }
-
-  return result;
-}

package/src/webperf/webperf-role-integration.js

@@ -1,66 +0,0 @@
-/**
- * Integration helpers that format CWV gate results for the
- * impeccable prompt and for pipeline progress events.
- */
-
-const RATING_ICONS = { good: "PASS", "needs-improvement": "WARN", poor: "FAIL" };
-
-/**
- * Build a human-readable section for the impeccable prompt.
- *
- * @param {{ pass: boolean, scores: object, blocking: Array, advisory: Array }} cwvResult
- * @returns {string}
- */
-export function buildWebPerfSection(cwvResult) {
-  const lines = [];
-  const verdict = cwvResult.pass ? "PASSED" : "FAILED";
-  lines.push(`## WebPerf — Core Web Vitals (${verdict})`);
-  lines.push("");
-
-  for (const [metric, { value, rating }] of Object.entries(cwvResult.scores)) {
-    const icon = RATING_ICONS[rating] || rating;
-    const unit = metric === "cls" ? "" : "ms";
-    lines.push(`- **${metric.toUpperCase()}**: ${value}${unit} [${icon}]`);
-  }
-
-  if (cwvResult.blocking.length > 0) {
-    lines.push("");
-    lines.push("### Blocking issues");
-    for (const b of cwvResult.blocking) {
-      const unit = b.metric === "cls" ? "" : "ms";
-      lines.push(`- ${b.metric.toUpperCase()}: ${b.value}${unit} exceeds poor threshold (${b.threshold}${unit})`);
-    }
-  }
-
-  if (cwvResult.advisory.length > 0) {
-    lines.push("");
-    lines.push("### Advisory (needs improvement)");
-    for (const a of cwvResult.advisory) {
-      const unit = a.metric === "cls" ? "" : "ms";
-      lines.push(`- ${a.metric.toUpperCase()}: ${a.value}${unit} exceeds good threshold (${a.threshold}${unit})`);
-    }
-  }
-
-  return lines.join("\n");
-}
-
-/**
- * Format CWV result for a pipeline progress event payload.
- *
- * @param {{ pass: boolean, scores: object, blocking: Array, advisory: Array }} cwvResult
- * @returns {{ type: string, pass: boolean, metrics: object, blockingCount: number, advisoryCount: number }}
- */
-export function formatCwvForEvent(cwvResult) {
-  const metrics = {};
-  for (const [metric, { value, rating }] of Object.entries(cwvResult.scores)) {
-    metrics[metric] = { value, rating };
-  }
-
-  return {
-    type: "webperf-cwv",
-    pass: cwvResult.pass,
-    metrics,
-    blockingCount: cwvResult.blocking.length,
-    advisoryCount: cwvResult.advisory.length
-  };
-}

package/templates/skills/kj-run.md

@@ -1,73 +0,0 @@
-# kj-run — Full Pipeline (Skills Mode)
-
-Execute the complete Karajan pipeline as sequential skills.
-
-## Your task
-
-$ARGUMENTS
-
-## Pipeline steps (execute in order)
-
-### Step 1 — Discover (optional but recommended)
-Analyze the task for gaps before coding:
-- Identify missing requirements, ambiguities, contradictions
-- If critical gaps found, STOP and ask the user before proceeding
-- If ready, continue
-
-### Step 2 — Code (with guardrails)
-Implement the task:
-1. **Tests first** (TDD): write/update tests before implementation
-2. **Implement**: minimal, focused code to fulfill the task
-3. **Verify**: run the test suite
-4. **Security check**: no hardcoded secrets, no injection vectors, no destructive ops in the diff
-5. **Diff check**: run `git diff` and verify only intended lines changed
-6. If any guardrail fails, fix before proceeding
-
-### Step 3 — Review (self-review)
-Review your own changes against quality standards:
-1. Run `git diff main...HEAD` (or base branch)
-2. Check: security, correctness, tests, architecture, style (in that order)
-3. Flag blocking issues:
-   - Hardcoded credentials or secrets
-   - Entire files overwritten instead of targeted edits
-   - Missing tests for new code
-   - SQL injection, XSS, command injection
-   - Destructive operations
-4. If blocking issues found, fix them and re-review
-5. If clean, proceed
-
-### Step 4 — Test audit
-Verify test quality:
-1. Every changed source file has corresponding tests
-2. Run `npm test` (or equivalent) — all must pass
-3. No skipped tests for changed code
-4. If tests fail, fix before proceeding
-
-### Step 5 — Security scan
-Quick security audit on the diff:
-1. Scan for OWASP top 10 in changed files
-2. Check for leaked secrets, injection vectors, missing auth
-3. If critical/high findings, fix before proceeding
-
-### Step 6 — Sonar (if available)
-If SonarQube is running (`docker ps | grep sonarqube`):
-1. Run `npx @sonar/scan`
-2. Check quality gate
-3. Fix blockers and critical issues
-
-### Step 7 — Commit
-If all steps pass:
-1. Stage changed files: `git add <specific files>`
-2. Commit with conventional commit message: `feat:`, `fix:`, `refactor:`, etc.
-3. Do NOT push unless the user explicitly asks
-
-## HU Board Integration
-If HU Board is enabled (`hu_board.enabled: true` in config), stories and sessions
-are automatically tracked and visible at the board URL. Run `kj board` to start it.
-
-## Important rules
-
-- **Never skip steps** — execute all applicable steps in order
-- **Fix before proceeding** — if a step finds issues, fix them before moving to the next
-- **Report progress** — after each step, briefly state what was done and the result
-- **Stop on critical** — if a critical security or correctness issue can't be fixed, stop and report