codeharness 0.30.1 → 0.31.0

package/README.md

@@ -8,7 +8,7 @@ codeharness is an **npm CLI** + **Claude Code plugin** that packages verificatio
 
 1. **Verifies features work** — not just that tests pass. Black-box verification runs the built CLI inside a Docker container with no source code access. If the feature doesn't work from a user's perspective, verification fails.
 2. **Fixes what it finds** — verification failures with code bugs automatically return to development with specific findings. The dev agent gets told exactly what's broken and why.
-3. **Runs sprints autonomously** — reads your sprint plan, picks the highest-priority story, implements it, reviews it, verifies it, and moves to the next one. Cross-epic prioritization, retry management, and session handoff built in.
+3. **Runs sprints autonomously** — reads your sprint plan, picks the highest-priority story, implements it, checks it (tests + lint), verifies it (agent evaluation), and moves to the next one. Cross-epic prioritization, retry management, and session handoff built in.
 4. **Makes agents see runtime** — ephemeral VictoriaMetrics stack (logs, metrics, traces) that agents query programmatically during development. No guessing at what the code does at runtime.
 
 ## Installation
@@ -61,7 +61,7 @@ The plugin provides slash commands that orchestrate the CLI within Claude Code s
 
 | Command | Purpose |
 |---------|---------|
-| `/harness-run` | Autonomous sprint execution — picks stories by priority, runs create → dev → review → verify loop |
+| `/harness-run` | Autonomous sprint execution — picks stories by priority, runs create → implement → check → verify loop |
 | `/harness-init` | Interactive project initialization |
 | `/harness-status` | Quick overview of sprint progress and harness health |
 | `/harness-onboard` | Scan project and generate onboarding plan |
@@ -84,7 +84,7 @@ codeharness integrates with [BMAD Method](https://github.com/bmadcode/BMAD-METHO
 ┌─────────────────────────────────────────┐
 │  Claude Code Session                     │
 │  /harness-run picks next story           │
-│  → create-story → dev → review → verify  │
+│  → create-story → implement → check → verify │
 └────────────────────┬────────────────────┘
                      │ verify
                      ▼

package/dist/{chunk-QLY7NJIB.js → chunk-ITPLJVAB.js}

@@ -2895,7 +2895,7 @@ function generateDockerfileTemplate(projectDir, stackOrDetections) {
 }
 
 // src/modules/infra/init-project.ts
-var HARNESS_VERSION = true ? "0.30.1" : "0.0.0-dev";
+var HARNESS_VERSION = true ? "0.31.0" : "0.0.0-dev";
 function failResult(opts, error) {
   return {
     status: "fail",

package/dist/{docker-P65B7Z3S.js → docker-TANMGEDO.js}

@@ -16,7 +16,7 @@ import {
   stopCollectorOnly,
   stopSharedStack,
   stopStack
-} from "./chunk-QLY7NJIB.js";
+} from "./chunk-ITPLJVAB.js";
 export {
   checkRemoteEndpoint,
   cleanupOrphanedContainers,

package/dist/index.js

@@ -40,7 +40,7 @@ import {
   validateDockerfile,
   warn,
   writeState
-} from "./chunk-QLY7NJIB.js";
+} from "./chunk-ITPLJVAB.js";
 
 // src/index.ts
 import { Command } from "commander";
@@ -5101,23 +5101,10 @@ import { Box as Box7, Static, Text as Text7, useInput } from "ink";
 // src/lib/ink-workflow.tsx
 import { Text as Text2, Box as Box2 } from "ink";
 import { jsx as jsx2, jsxs as jsxs2 } from "react/jsx-runtime";
-var termWidth = () => Math.min(process.stdout.columns || 60, 80);
 var SPINNER_FRAMES = ["\u280B", "\u2819", "\u2839", "\u2838", "\u283C", "\u2834", "\u2826", "\u2827", "\u2807", "\u280F"];
 function isLoopBlock2(step) {
   return typeof step === "object" && step !== null && "loop" in step;
 }
-function formatCost(costUsd) {
-  if (costUsd == null) return "...";
-  return `$${costUsd.toFixed(2)}`;
-}
-function formatElapsed2(ms) {
-  if (ms == null) return "...";
-  const seconds = Math.round(ms / 1e3);
-  if (seconds >= 60) {
-    return `${Math.floor(seconds / 60)}m`;
-  }
-  return `${seconds}s`;
-}
 function TaskNode({ name, status, spinnerFrame }) {
   const s = status ?? "pending";
   switch (s) {
@@ -5151,17 +5138,6 @@ function loopIteration(tasks, taskStates) {
   });
   return anyStarted ? 1 : 0;
 }
-function collectTaskNames(flow) {
-  const names = [];
-  for (const step of flow) {
-    if (isLoopBlock2(step)) {
-      names.push(...step.loop);
-    } else {
-      names.push(step);
-    }
-  }
-  return names;
-}
 function hasMetaData(taskMeta) {
   if (!taskMeta) return false;
   return Object.keys(taskMeta).length > 0;
@@ -5207,69 +5183,10 @@ function WorkflowGraph({ flow, currentTask, taskStates, taskMeta }) {
       );
     }
   }
-  let driverRow = null;
-  let costRow = null;
-  if (showMeta) {
-    const taskNames = collectTaskNames(flow);
-    const driverParts = [];
-    const costParts = [];
-    let hasAnyCost = false;
-    for (const name of taskNames) {
-      const m = meta[name];
-      const driver = m?.driver ?? "";
-      driverParts.push(driver);
-      const state = taskStates[name];
-      if (state === "done") {
-        const costStr = formatCost(m?.costUsd);
-        const timeStr = formatElapsed2(m?.elapsedMs);
-        costParts.push(`${costStr} / ${timeStr}`);
-        hasAnyCost = true;
-      } else {
-        costParts.push("");
-      }
-    }
-    const hasSomeDriver = driverParts.some((d) => d.length > 0);
-    if (hasSomeDriver) {
-      const driverLabels = [];
-      for (let idx = 0; idx < taskNames.length; idx++) {
-        if (idx > 0) {
-          driverLabels.push(/* @__PURE__ */ jsx2(Text2, { children: "   " }, `drv-sep-${idx}`));
-        }
-        driverLabels.push(
-          /* @__PURE__ */ jsx2(Text2, { dimColor: true, children: driverParts[idx] || " " }, `drv-${idx}`)
-        );
-      }
-      driverRow = /* @__PURE__ */ jsxs2(Text2, { children: [
-        "  ",
-        driverLabels
-      ] });
-    }
-    if (hasAnyCost) {
-      const costLabels = [];
-      for (let idx = 0; idx < taskNames.length; idx++) {
-        if (idx > 0) {
-          costLabels.push(/* @__PURE__ */ jsx2(Text2, { children: "   " }, `cost-sep-${idx}`));
-        }
-        costLabels.push(
-          /* @__PURE__ */ jsx2(Text2, { dimColor: true, children: costParts[idx] || " " }, `cost-${idx}`)
-        );
-      }
-      costRow = /* @__PURE__ */ jsxs2(Text2, { children: [
-        "  ",
-        costLabels
-      ] });
-    }
-  }
-  return /* @__PURE__ */ jsxs2(Box2, { flexDirection: "column", children: [
-    /* @__PURE__ */ jsx2(Text2, { children: "\u2501".repeat(termWidth()) }),
-    /* @__PURE__ */ jsxs2(Text2, { children: [
-      "  ",
-      elements
-    ] }),
-    driverRow,
-    costRow,
-    /* @__PURE__ */ jsx2(Text2, { children: "\u2501".repeat(termWidth()) })
-  ] });
+  return /* @__PURE__ */ jsx2(Box2, { flexDirection: "column", children: /* @__PURE__ */ jsxs2(Text2, { children: [
+    "  ",
+    elements
+  ] }) });
 }
 
 // src/lib/ink-lane-container.tsx
@@ -5447,7 +5364,7 @@ import { jsx as jsx5, jsxs as jsxs5 } from "react/jsx-runtime";
 function formatConflictText(count) {
   return count === 1 ? "1 conflict" : `${count} conflicts`;
 }
-function formatCost2(cost) {
+function formatCost(cost) {
   return `$${cost.toFixed(2)}`;
 }
 function SummaryBar({ doneStories, mergingEpic, pendingEpics, completedLanes }) {
@@ -5473,7 +5390,7 @@ function SummaryBar({ doneStories, mergingEpic, pendingEpics, completedLanes })
       /* @__PURE__ */ jsx5(Text5, { children: " \u2502 " }),
       /* @__PURE__ */ jsx5(Text5, { dimColor: true, children: `Pending: ${pendingSection}` })
     ] }),
-    completedLanes && completedLanes.length > 0 && completedLanes.map((lane) => /* @__PURE__ */ jsx5(Text5, { color: "green", children: `[OK] Lane ${lane.laneIndex}: Epic ${lane.epicId} complete (${lane.storyCount} stories, ${formatCost2(lane.cost)}, ${lane.elapsed})` }, `lane-complete-${lane.laneIndex}`))
+    completedLanes && completedLanes.length > 0 && completedLanes.map((lane) => /* @__PURE__ */ jsx5(Text5, { color: "green", children: `[OK] Lane ${lane.laneIndex}: Epic ${lane.epicId} complete (${lane.storyCount} stories, ${formatCost(lane.cost)}, ${lane.elapsed})` }, `lane-complete-${lane.laneIndex}`))
   ] });
 }
 
@@ -5610,7 +5527,7 @@ function shortKey(key) {
   const m = key.match(/^(\d+-\d+)/);
   return m ? m[1] : key;
 }
-function formatCost3(cost) {
+function formatCost2(cost) {
   return `$${cost.toFixed(2)}`;
 }
 function Header({ info: info3, laneCount }) {
@@ -5619,7 +5536,7 @@ function Header({ info: info3, laneCount }) {
   if (laneCount != null && laneCount > 1) parts.push(`${laneCount} lanes`);
   if (info3.elapsed) parts.push(`${info3.elapsed} elapsed`);
   const displayCost = laneCount != null && laneCount > 1 && info3.laneTotalCost != null ? info3.laneTotalCost : info3.totalCost;
-  if (displayCost != null) parts.push(`${formatCost3(displayCost)} spent`);
+  if (displayCost != null) parts.push(`${formatCost2(displayCost)} spent`);
   const left = parts.join(" | ");
   const right = "[q to quit]";
   const width = process.stdout.columns || 80;
@@ -10559,7 +10476,7 @@ async function handleDockerCheck(isJson) {
     }
   }
 }
-function formatElapsed3(ms) {
+function formatElapsed2(ms) {
   const s = Math.floor(ms / 1e3);
   const h = Math.floor(s / 3600);
   const m = Math.floor(s % 3600 / 60);
@@ -10579,7 +10496,7 @@ function printWorkflowState() {
   console.log(`  Tasks completed: ${state.tasks_completed.length}`);
   if (state.phase === "executing" && state.started) {
     const elapsed = Date.now() - Date.parse(state.started);
-    console.log(`  Elapsed: ${formatElapsed3(elapsed)}`);
+    console.log(`  Elapsed: ${formatElapsed2(elapsed)}`);
   }
   if (state.evaluator_scores.length > 0) {
     const latest = state.evaluator_scores[state.evaluator_scores.length - 1];
@@ -10604,7 +10521,7 @@ function getWorkflowStateData() {
   };
   if (state.phase === "executing" && state.started) {
     data.elapsed_ms = Date.now() - Date.parse(state.started);
-    data.elapsed = formatElapsed3(data.elapsed_ms);
+    data.elapsed = formatElapsed2(data.elapsed_ms);
   }
   return data;
 }
@@ -11256,7 +11173,7 @@ function registerTeardownCommand(program) {
     } else if (otlpMode === "remote-routed") {
       if (!options.keepDocker) {
         try {
-          const { stopCollectorOnly: stopCollectorOnly2 } = await import("./docker-P65B7Z3S.js");
+          const { stopCollectorOnly: stopCollectorOnly2 } = await import("./docker-TANMGEDO.js");
           stopCollectorOnly2();
           result.docker.stopped = true;
           if (!isJson) {
@@ -11288,7 +11205,7 @@ function registerTeardownCommand(program) {
         info("Shared stack: kept running (other projects may use it)");
       }
     } else if (isLegacyStack) {
-      const { isStackRunning: isStackRunning2, stopStack } = await import("./docker-P65B7Z3S.js");
+      const { isStackRunning: isStackRunning2, stopStack } = await import("./docker-TANMGEDO.js");
       let stackRunning = false;
       try {
         stackRunning = isStackRunning2(composeFile);
@@ -13829,12 +13746,12 @@ var CodexDriver = class {
         opts.plugins
       );
     }
-    const args = [];
+    const args = ["exec"];
     if (opts.model) {
       args.push("--model", opts.model);
     }
     if (opts.cwd) {
-      args.push("--cwd", opts.cwd);
+      args.push("--cd", opts.cwd);
     }
     args.push(opts.prompt);
     let yieldedResult = false;
@@ -14184,7 +14101,7 @@ function registerDriversCommand(program) {
 }
 
 // src/index.ts
-var VERSION = true ? "0.30.1" : "0.0.0-dev";
+var VERSION = true ? "0.31.0" : "0.0.0-dev";
 function createProgram() {
   const program = new Command();
   program.name("codeharness").description("Makes autonomous coding agents produce software that actually works").version(VERSION).option("--json", "Output in machine-readable JSON format");

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "codeharness",
-  "version": "0.30.1",
+  "version": "0.31.0",
   "type": "module",
   "description": "CLI for codeharness — makes autonomous coding agents produce software that actually works",
   "bin": {

package/templates/agents/checker.yaml

@@ -0,0 +1,65 @@
+name: checker
+role:
+  title: Automated Checker
+  purpose: Run tests, linter, and coverage checks — report pass/fail objectively
+persona:
+  identity: |
+    CI bot that runs the project's test suite, linter, and coverage tool.
+    Reports results objectively — no interpretation, no fixes, just facts.
+  communication_style: "Machine-like. Commands run, output captured, pass/fail reported."
+  principles:
+    - Run the project's actual test command (npm test, pytest, cargo test, etc.)
+    - Run the project's linter if configured (eslint, ruff, clippy, etc.)
+    - Check coverage against target if configured
+    - Report exact command, exit code, and output for each check
+    - Never fix code — only report results
+prompt_template: |
+  ## Role
+
+  You are running automated checks on the implementation. Run tests, linter, and coverage. Report results.
+
+  ## Process
+
+  1. **Detect check commands** from the project (package.json scripts, pyproject.toml, Makefile, etc.)
+  2. **Run tests**: execute the test command, capture output and exit code
+  3. **Run linter**: execute the lint command if available
+  4. **Check coverage**: if a coverage target exists, verify it's met
+
+  ## Output Format
+
+  Output a single JSON object:
+
+  ```json
+  {
+    "verdict": "pass" | "fail",
+    "checks": [
+      {
+        "name": "tests",
+        "command": "npm test",
+        "exit_code": 0,
+        "passed": true,
+        "summary": "42 tests passed"
+      },
+      {
+        "name": "lint",
+        "command": "npm run lint",
+        "exit_code": 0,
+        "passed": true,
+        "summary": "no issues"
+      },
+      {
+        "name": "coverage",
+        "command": "npm run coverage",
+        "exit_code": 0,
+        "passed": true,
+        "summary": "98% (target: 100%)"
+      }
+    ]
+  }
+  ```
+
+  Verdict is "pass" only if ALL checks pass.
+
+  ## Output Location
+
+  Write results to ./verdict/check.json

package/templates/workflows/default.yaml

@@ -11,6 +11,12 @@ tasks:
     session: fresh
     source_access: true
     model: claude-sonnet-4-6
+  check:
+    agent: checker
+    scope: per-story
+    session: fresh
+    source_access: true
+    driver: codex
   review:
     agent: reviewer
     scope: per-story
@@ -39,10 +45,12 @@ tasks:
 flow:
   - create-story
   - implement
+  - check
   - review
   - verify
   - loop:
       - retry
+      - check
       - review
       - verify
   - retro