patchwarden 1.1.0 → 1.5.1

package/README.en.md

@@ -8,7 +8,7 @@
 [![Node.js](https://img.shields.io/badge/Node.js-%3E%3D18-339933.svg)](https://nodejs.org/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
 
-Current source version: **v1.1.0**. See the
+Current source version: **v1.5.1**. See the
 [CHANGELOG](CHANGELOG.md), [migration guide](docs/migration-from-safe-bifrost.md), and
 [release checklist](docs/release-checklist.md). Verify GitHub Release / npm publication separately before release.
 
@@ -46,6 +46,7 @@ independent verification evidence.
 - [HTTP MCP mode](#http-mcp-mode)
 - [Diagnostics and health checks](#diagnostics-and-health-checks)
 - [Lessons learned and troubleshooting](#lessons-learned-and-troubleshooting)
+- [Ecosystem guides](#ecosystem-guides)
 - [Security boundaries and local data](#security-boundaries-and-local-data)
 - [Upgrading and migration](#upgrading-and-migration)
 - [Development and release verification](#development-and-release-verification)
@@ -493,7 +494,7 @@ PatchWarden.cmd start core
 The launcher:
 
 - Builds `dist/index.js` if it is missing.
-- Verifies v1.1.0, the fixed 21-tool `chatgpt_core` catalog, and its schema
+- Verifies v1.5.0, the fixed 26-tool `chatgpt_core` catalog, and its schema
   manifest.
 - Reads or prompts for the Tunnel ID.
 - Reads or prompts for the runtime API key.
@@ -845,6 +846,31 @@ It stops only this project's launchers and Watcher plus `tunnel-client.exe`
 processes that exactly match the selected profile. It does not globally
 terminate unrelated PatchWarden, OpenCode, or Codex instances.
 
+## Dashboard pages and recommended workflow
+
+The PatchWarden Dashboard (Control Center) provides these main pages:
+
+- **Dashboard**: System overview, including Repo selector, Health Score, service status, Release card, Project Policy, Lineage, Evidence Pack, stale task hints, recent task list, and system status (with Copy diagnostics).
+- **Tasks**: Task list, filterable by repo_path / status / acceptance_status / warning type / agent / date range. Each row offers safe_result / safe_audit / safe_test_summary / safe_diff_summary quick actions.
+- **Task Detail**: Defaults to a safe summary (safe_result / safe_test_summary / safe_diff_summary / safe_audit). Full result / diff / test log live in a collapsed advanced section and load on demand.
+- **Direct Sessions**: Grouped by active / finalized / audited / expired, with safe_direct_summary / safe_finalize_direct_session / safe_audit_direct_session quick actions.
+- **Audit / Warnings**: Aggregated by warning type, showing affected tasks / severity / recommended action.
+- **Workspace**: First-level workspace directories and project list.
+- **Logs**: Tails of Core / Direct / Watcher / Control Center logs.
+
+Recommended workflow:
+
+1. Select the target repo at the top of the Dashboard.
+2. Check the Health Score to confirm system health.
+3. Click safe_result on a recent task row to view the task summary.
+4. Open Task Detail for safe-first acceptance; expand the advanced section only when you need the full logs.
+5. Use Lineage Detail to inspect why a run_task_loop succeeded or failed.
+6. Export the Evidence Pack after acceptance.
+7. Use Direct sessions for independent verification, then finalize + audit.
+8. When something goes wrong, click Copy diagnostics and paste the output to ChatGPT / Codex / opencode for triage.
+
+> See [docs/dashboard-overview.md](docs/dashboard-overview.md), [docs/task-safe-review-workflow.md](docs/task-safe-review-workflow.md), [docs/lineage-evidence-pack-workflow.md](docs/lineage-evidence-pack-workflow.md), and [docs/direct-session-workflow.md](docs/direct-session-workflow.md) for details.
+
 ## Lessons learned and troubleshooting
 
 ### Quick reference
@@ -930,10 +956,12 @@ names.
 
 ## MCP tools and profiles
 
-`chatgpt_core` is the fixed 21-tool profile used by the ChatGPT tunnel:
+`chatgpt_core` is the fixed 26-tool profile used by the ChatGPT tunnel:
 
 `health_check`, `list_agents`, `list_workspace`,
-`read_workspace_file`, `save_plan`, `create_task`,
+`read_workspace_file`, `save_plan`, `create_task`, `run_task_loop`,
+`recommend_agent_for_task`, `get_task_lineage`, `export_task_evidence_pack`,
+`get_project_policy`,
 `wait_for_task`, `get_task_summary`, `get_diff`, `get_result`,
 `get_result_json`, `get_test_log`, `get_task_status`, `list_tasks`,
 `cancel_task`, `audit_task`, `safe_status`, `safe_result`, `safe_audit`, `safe_test_summary`, and `safe_diff_summary`.
@@ -942,6 +970,38 @@ names.
 ChatGPT should request `view: "compact"` first; terminal `wait_for_task`
 responses also embed compact acceptance evidence only.
 
+`run_task_loop` is the v1.2 safe orchestration entrypoint. It only composes
+existing `create_task`, `wait_for_task`, safe summary, and `audit_task`
+behavior; it does not bypass the Watcher, command allowlist, workspace
+confinement, or local confirmation boundaries. `get_task_lineage` reads the
+bounded `.patchwarden/lineages/<lineage_id>/` summary without returning full
+logs or diffs.
+
+`get_project_policy` is the v1.3 read-only policy entrypoint. It returns the
+bounded effective `.patchwarden/project-policy.json` policy and release
+readiness summary without expanding command permissions. The v1.3 release mode
+tools, `release_check`, `release_prepare`, `release_verify`, and
+`release_cleanup`, are full-profile only and never perform publish, push, tag,
+or GitHub Release writes. The Control Center dashboard also shows bounded
+lineage, policy, and release status summaries without full logs, diffs, or
+secret-bearing content.
+
+v1.4 adds Direct-assisted loop verification. `run_task_loop` can opt into
+`direct_verify=true`; after the guarded task and audit succeed, PatchWarden
+creates a Direct session, runs only allowlisted verification commands, safe
+finalizes, safe audits, and records bounded Direct evidence in lineage. It does
+not call Direct patching tools, publish, push, tag, create releases, restart
+watchers, or return full stdout/stderr/diffs.
+
+v1.5 adds worktree-assisted loop isolation, bounded agent routing, and evidence
+pack export. `run_task_loop(isolation_mode="worktree")` creates an isolated git
+worktree for the task and records the worktree id, path, branch, and next action
+in lineage. It does not auto-merge or auto-delete the worktree. `agent="auto"`
+uses `recommend_agent_for_task` to select a configured agent before task
+creation. `export_task_evidence_pack` writes bounded `evidence.json` and
+`EVIDENCE.md` files under `.patchwarden/evidence-packs/<lineage_id>/` without
+stdout/stderr tails, full diffs, verification logs, or sensitive file content.
+
 `full` additionally provides:
 
 - `get_plan`
@@ -956,9 +1016,9 @@ to `full`.
 
 ### ChatGPT Direct mode
 
-Direct mode exposes thirteen guarded tools so ChatGPT can create an editing
+Direct mode exposes fourteen guarded tools so ChatGPT can create an editing
 session, read and search source files, apply hash-bound JSON patches, run
-exactly allowlisted verification commands, finalize the evidence, and audit
+exactly allowlisted verification commands or a bounded verification bundle, finalize the evidence, and audit
 the result without a local execution agent.
 
 Enable it in the trusted local configuration while keeping the ordinary Core
@@ -980,9 +1040,20 @@ On first use, provide the `tunnel-client.exe` path and a Tunnel ID dedicated
 to the Direct Connector. The launcher uses the `patchwarden-direct` profile,
 stores runtime state under `%LOCALAPPDATA%\patchwarden\runtime-direct`, skips
 the Watcher, and retains the existing DPAPI credential handling. In a fresh
-ChatGPT conversation, `health_check` should report `chatgpt_direct`, thirteen
+ChatGPT conversation, `health_check` should report `chatgpt_direct`, fourteen
 tools, and `direct_profile_enabled=true`.
 
+## Ecosystem guides
+
+- [Why PatchWarden](docs/why-patchwarden.md)
+- [Evidence Pack v2 schema](docs/evidence-pack-schema.md)
+- [Spec Kit integration](docs/spec-kit-integration.md)
+- [AgentSeal integration](docs/agentseal-integration.md)
+- [MCP Inspector testing](docs/mcp-inspector-testing.md)
+- [OpenCode worker integration](docs/opencode-worker.md)
+- [OpenHands worker integration](docs/openhands-worker.md)
+- [Threat model](docs/threat-model.md)
+
 ## Security boundaries and local data
 
 Primary protections:
@@ -1088,6 +1159,10 @@ and release-asset checksums independently.
 - [ChatGPT usage guide](docs/chatgpt-usage.md)
 - [Migration guide](docs/migration-from-safe-bifrost.md)
 - [ChatGPT Connector demo](docs/demo.md)
+- [Dashboard overview](docs/dashboard-overview.md)
+- [Task safe review workflow](docs/task-safe-review-workflow.md)
+- [Lineage and Evidence Pack workflow](docs/lineage-evidence-pack-workflow.md)
+- [Direct session workflow](docs/direct-session-workflow.md)
 - [OpenAI tunnel examples](examples/openai-tunnel/README.md)
 - [ChatGPT test prompt](examples/openai-tunnel/chatgpt-test-prompt.md)
 
@@ -1103,7 +1178,7 @@ and release-asset checksums independently.
 - [x] Release Gate (five-stage pre-release verification)
 - [x] Worktree isolation
 - [x] Multi-agent routing
-- [ ] Local dashboard
+- [x] Local dashboard
 
 ## License
 

package/README.md

@@ -8,7 +8,7 @@
 [![Node.js](https://img.shields.io/badge/Node.js-%3E%3D18-339933.svg)](https://nodejs.org/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
 
-当前源码版本：**v1.1.0**。查看
+当前源码版本：**v1.5.1**。查看
 [CHANGELOG](CHANGELOG.md)、[迁移指南](docs/migration-from-safe-bifrost.md)和
 [发布检查清单](docs/release-checklist.md)。GitHub Release / npm 发布状态需要在发布前单独核对。
 
@@ -44,6 +44,7 @@ PatchWarden 负责把计划保存成工作区内任务，再由预先配置的
 - [诊断与健康检查](#诊断与健康检查)
 - [踩坑记录与故障排查](#踩坑记录与故障排查)
 - [Direct 模式：ChatGPT 直接开发](#direct-模式chatgpt-直接开发)
+- [生态适配](#生态适配)
 - [安全边界与本地数据](#安全边界与本地数据)
 - [升级与旧版本迁移](#升级与旧版本迁移)
 - [开发与发布验证](#开发与发布验证)
@@ -399,7 +400,7 @@ PatchWarden.cmd start core
 启动器会：
 
 - 检查 `dist/index.js`，缺失时自动构建。
-- 校验 `chatgpt_core` 的版本、21 个核心工具和 Schema Manifest。
+- 校验 `chatgpt_core` 的版本、26 个核心工具和 Schema Manifest。
 - 读取或提示输入 Tunnel ID。
 - 读取或提示输入运行时 API Key。
 - 使用 Windows DPAPI 保存凭据到 `%APPDATA%\patchwarden`。
@@ -751,6 +752,31 @@ powershell.exe -NoProfile -ExecutionPolicy Bypass -File "%CD%\scripts\control\st
 `chatgpt_core` / `patchwarden` → `127.0.0.1:8080`，
 `chatgpt_direct` / `patchwarden-direct` → `127.0.0.1:8081`。
 
+## Dashboard 主要页面与推荐工作流
+
+PatchWarden Dashboard (Control Center) 提供以下主要页面：
+
+- **控制台 (Dashboard)**: 系统总览，包含 Repo selector、Health Score、服务状态、Release 卡片、Project Policy、Lineage、Evidence Pack、Stale 任务提示、最近任务列表、系统状态（含 Copy diagnostics）。
+- **任务面板 (Tasks)**: 任务列表，支持按 repo_path / status / acceptance_status / warning type / agent / date range 过滤，每行提供 safe_result / safe_audit / safe_test_summary / safe_diff_summary 快捷操作。
+- **任务详情 (Task Detail)**: 默认展示 safe 摘要（safe_result / safe_test_summary / safe_diff_summary / safe_audit），完整 result / diff / test log 在折叠的高级区按需加载。
+- **Direct 会话 (Direct Sessions)**: 按 active / finalized / audited / expired 分组，提供 safe_direct_summary / safe_finalize_direct_session / safe_audit_direct_session 快捷操作。
+- **审计日志 / Warnings (Audit)**: 按 warning 类型聚合，显示 affected tasks / severity / recommended action。
+- **工作区 (Workspace)**: workspace 一级目录与项目列表。
+- **日志 (Logs)**: Core / Direct / Watcher / Control Center 日志尾部。
+
+推荐工作流：
+
+1. 在 Dashboard 顶部选择目标 repo。
+2. 查看 Health Score 确认系统健康。
+3. 在最近任务列表点击 safe_result 快捷查看任务摘要。
+4. 打开 Task Detail 进行 safe-first 验收，仅在需要时展开高级区查看完整日志。
+5. 使用 Lineage Detail 查看 run_task_loop 的成功/失败原因。
+6. 验收完成后导出 Evidence Pack。
+7. Direct 会话用于独立验证，完成后 finalize + audit。
+8. 遇到问题时点击 Copy diagnostics 复制诊断信息发给 ChatGPT / Codex / opencode 排查。
+
+> 详细说明见 [docs/dashboard-overview.md](docs/dashboard-overview.md)、[docs/task-safe-review-workflow.md](docs/task-safe-review-workflow.md)、[docs/lineage-evidence-pack-workflow.md](docs/lineage-evidence-pack-workflow.md)、[docs/direct-session-workflow.md](docs/direct-session-workflow.md)。
+
 ## 踩坑记录与故障排查
 
 ### 快速对照表
@@ -836,10 +862,11 @@ PatchWarden v0.4.0 不自动读取旧 CLI、旧环境变量、旧 Header、旧
 
 ## MCP 工具与 Profile
 
-`chatgpt_core` 是固定的 21 工具 Profile，适合 ChatGPT Tunnel：
+`chatgpt_core` 是固定的 26 工具 Profile，适合 ChatGPT Tunnel：
 
 `health_check`、`list_agents`、`list_workspace`、
-`read_workspace_file`、`save_plan`、`create_task`、
+`read_workspace_file`、`save_plan`、`create_task`、`run_task_loop`、`recommend_agent_for_task`、
+`get_task_lineage`、`export_task_evidence_pack`、`get_project_policy`、
 `wait_for_task`、`get_task_summary`、`get_diff`、`get_result`、
 `get_result_json`、`get_test_log`、`get_task_status`、`list_tasks`、
 `cancel_task`、`audit_task`、`safe_status`、`safe_result`、`safe_audit`、`safe_test_summary`、`safe_diff_summary`。
@@ -847,7 +874,33 @@ PatchWarden v0.4.0 不自动读取旧 CLI、旧环境变量、旧 Header、旧
 `get_task_summary` 默认保留兼容的 `standard` 视图；ChatGPT 应优先使用
 `view: "compact"`，终态 `wait_for_task` 也只内嵌 compact 验收证据。
 
-`full` 提供完整 54 工具本地开发目录，包含核心工具、管理工具和 Direct
+`run_task_loop` 是 v1.2 的安全编排入口：它只组合现有 `create_task`、`wait_for_task`、safe summary 和
+`audit_task`，不会绕过 Watcher、命令白名单、workspace confinement 或确认边界。`get_task_lineage`
+读取 `.patchwarden/lineages/<lineage_id>/` 中的有界链路摘要，不返回完整日志或 diff。
+
+v1.4 adds Direct-assisted loop verification. `run_task_loop(direct_verify=true)`
+will create a Direct session only after the watcher-driven task and normal audit
+succeed, run allowlisted Direct verification commands, safe-finalize, safe-audit,
+and write bounded Direct evidence into lineage. It does not call Direct patching
+tools, publish, push, tag, create releases, restart live services, or return full
+stdout/stderr/diffs.
+
+v1.5 adds worktree-assisted loop isolation, bounded agent routing, and evidence
+pack export. `run_task_loop(isolation_mode="worktree")` creates an isolated git
+worktree for the task and records the worktree id, path, branch, and next action
+in lineage. It does not auto-merge or auto-delete the worktree. `agent="auto"`
+uses `recommend_agent_for_task` to select a configured agent before task
+creation. `export_task_evidence_pack` writes bounded `evidence.json` and
+`EVIDENCE.md` files under `.patchwarden/evidence-packs/<lineage_id>/` without
+stdout/stderr tails, full diffs, verification logs, or sensitive file content.
+
+`get_project_policy` 是 v1.3 的只读策略入口，返回 `.patchwarden/project-policy.json`
+的有界 effective policy 与 release readiness，不会扩大命令白名单。v1.3 的
+`release_check`、`release_prepare`、`release_verify`、`release_cleanup` 仅在 `full`
+Profile 中可见，不执行 publish、push、tag 或 GitHub Release 写操作。Control Center
+Dashboard 会展示 lineage、policy 和 release status 的有界摘要，不返回完整日志、diff 或密钥内容。
+
+`full` 提供完整 64 工具本地开发目录，包含核心工具、管理工具和 Direct
 工具。除 `chatgpt_core` 外，常用额外管理工具包括：
 
 - `get_plan`
@@ -857,15 +910,15 @@ PatchWarden v0.4.0 不自动读取旧 CLI、旧环境变量、旧 Header、旧
 - `get_task_stdout_tail`
 - `get_task_log_tail`
 
-`chatgpt_direct` 是 v0.6.0 新增的 Direct 直接开发 Profile，v1.1 包含 13 个工具：
+`chatgpt_direct` 是 v0.6.0 新增的 Direct 直接开发 Profile，v1.4 包含 14 个工具：
 
 `health_check`、`list_workspace`、`create_direct_session`、
 `search_workspace`、`read_workspace_file`、`apply_patch`、
-`run_verification`、`finalize_direct_session`、`audit_session`、`safe_direct_summary`、`safe_finalize_direct_session`、`safe_audit_direct_session`、`sync_file`。
+`run_verification`、`run_direct_verification_bundle`、`finalize_direct_session`、`audit_session`、`safe_direct_summary`、`safe_finalize_direct_session`、`safe_audit_direct_session`、`sync_file`。
 
 `chatgpt_direct` 默认关闭，需要通过 `enableDirectProfile: true` 或
 `PATCHWARDEN_TOOL_PROFILE=chatgpt_direct` 显式启用。`chatgpt_core` 保持
-21 工具清单。
+26 工具清单。
 
 Tunnel 包装脚本会强制使用 `chatgpt_core`；普通本地开发默认使用 `full`。
 
@@ -920,7 +973,7 @@ Tunnel ID。启动器使用 `patchwarden-direct` Profile、独立的
 Tunnel API Key 仍通过现有 Windows DPAPI 缓存处理，不会写入仓库。
 
 连接 ChatGPT 后新建对话并先调用 `health_check`。预期
-`tool_profile=chatgpt_direct`、`tool_count=13`、
+`tool_profile=chatgpt_direct`、`tool_count=14`、
 `direct_profile_enabled=true`。已有对话可能缓存旧工具清单，需要重新连接
 Connector 后再新建对话。
 
@@ -950,6 +1003,17 @@ Direct 模式在受控 session 内执行，具有以下硬性限制：
 
 Direct 模式不支持：shell、delete、rename、commit、push、publish、deployment。
 
+## 生态适配
+
+- [为什么需要 PatchWarden](docs/why-patchwarden.md)
+- [Evidence Pack v2 文件结构](docs/evidence-pack-schema.md)
+- [Spec Kit 集成模式](docs/spec-kit-integration.md)
+- [AgentSeal 集成模式](docs/agentseal-integration.md)
+- [MCP Inspector 测试](docs/mcp-inspector-testing.md)
+- [OpenCode worker 集成](docs/opencode-worker.md)
+- [OpenHands worker 集成](docs/openhands-worker.md)
+- [威胁模型](docs/threat-model.md)
+
 ## 安全边界与本地数据
 
 PatchWarden 的主要保护：
@@ -1054,6 +1118,10 @@ Release 和发布资产校验值。
 - [ChatGPT 调用规范](docs/chatgpt-usage.md)
 - [旧版本迁移指南](docs/migration-from-safe-bifrost.md)
 - [ChatGPT Connector 演示](docs/demo.md)
+- [Dashboard 概览](docs/dashboard-overview.md)
+- [任务 safe 验收工作流](docs/task-safe-review-workflow.md)
+- [Lineage 与 Evidence Pack 工作流](docs/lineage-evidence-pack-workflow.md)
+- [Direct 会话工作流](docs/direct-session-workflow.md)
 - [OpenAI Tunnel 示例](examples/openai-tunnel/README.md)
 - [ChatGPT 测试提示词](examples/openai-tunnel/chatgpt-test-prompt.md)
 
@@ -1069,7 +1137,7 @@ Release 和发布资产校验值。
 - [x] Release Gate（发布前五阶段校验）
 - [x] Worktree 隔离
 - [x] 多 Agent 路由
-- [ ] 本地 Dashboard
+- [x] 本地 Dashboard
 
 ## License
 

package/dist/assessments/agentAssessor.d.ts

@@ -12,4 +12,4 @@ export interface AgentAssessorInput {
     maxOutputBytes: number;
     config: PatchWardenConfig;
 }
-export declare function runAgentAssessment(input: AgentAssessorInput): AgentAssessmentSummary;
+export declare function runAgentAssessment(input: AgentAssessorInput): Promise<AgentAssessmentSummary>;

package/dist/assessments/agentAssessor.js

@@ -4,7 +4,7 @@ import { buildAgentInvocation, buildAssessmentPrompt } from "../runner/agentInvo
 import { runSimpleProcessSync } from "../runner/simpleProcess.js";
 import { captureRepoSnapshot, compareSnapshots } from "../runner/changeCapture.js";
 const ASSESSMENT_MARKER = "===ASSESSMENT_JSON===";
-export function runAgentAssessment(input) {
+export async function runAgentAssessment(input) {
     const logPaths = {
         stdout: resolve(input.assessmentDir, "agent-assessment-stdout.log"),
         stderr: resolve(input.assessmentDir, "agent-assessment-stderr.log"),
@@ -33,7 +33,7 @@ export function runAgentAssessment(input) {
     // ── 2. Before snapshot (repo-scoped) ──
     let repoBefore;
     try {
-        repoBefore = captureRepoSnapshot(input.repoPath);
+        repoBefore = await captureRepoSnapshot(input.repoPath);
     }
     catch {
         // If we can't capture before snapshot, conservatively skip agent assessment
@@ -73,7 +73,7 @@ export function runAgentAssessment(input) {
     let readOnlyViolation = false;
     let violationFiles = [];
     try {
-        const repoAfter = captureRepoSnapshot(input.repoPath);
+        const repoAfter = await captureRepoSnapshot(input.repoPath);
         const changes = compareSnapshots(repoBefore, repoAfter);
         if (changes.length > 0) {
             readOnlyViolation = true;

package/dist/assessments/assessmentStore.d.ts

@@ -117,7 +117,7 @@ export declare function validateAssessmentFreshness(assessmentId: string, curren
  * This is intentionally not registered as an MCP tool: a remote client may
  * request confirmation, but it cannot grant confirmation to itself.
  */
-export declare function confirmAssessment(assessmentId: string): AssessmentConfirmationResult;
+export declare function confirmAssessment(assessmentId: string): Promise<AssessmentConfirmationResult>;
 export declare function computeWorkspaceFingerprint(snapshot: RepoSnapshot): string;
 interface PolicyHashInput {
     change_policy: ChangePolicy;

package/dist/assessments/assessmentStore.js

@@ -156,7 +156,7 @@ export function validateAssessmentFreshness(assessmentId, currentSnapshot, optio
  * This is intentionally not registered as an MCP tool: a remote client may
  * request confirmation, but it cannot grant confirmation to itself.
  */
-export function confirmAssessment(assessmentId) {
+export async function confirmAssessment(assessmentId) {
     if (!/^assessment_\d{8}_\d{6}_[0-9a-f]{32}$/.test(assessmentId)) {
         throw new PatchWardenError("assessment_id_invalid", "Local confirmation requires the full assessment_id (32 hexadecimal random characters).", "Copy the complete assessment_id from the assess_only response; assessment_short_id is display-only.");
     }
@@ -166,7 +166,7 @@ export function confirmAssessment(assessmentId) {
             ? "Fix the hard-rule finding and run assess_only again."
             : "This assessment does not require confirmation; execute it with the minimal assessment_id call.");
     }
-    const snapshot = captureRepoSnapshot(assessment.resolved_repo_path);
+    const snapshot = await captureRepoSnapshot(assessment.resolved_repo_path);
     const validation = validateAssessmentFreshness(assessmentId, snapshot, {
         allow_unconfirmed: true,
         // A standalone CLI has no active MCP profile snapshot. The execute path

package/dist/assessments/confirmCli.js

@@ -9,18 +9,19 @@
 import { loadConfig } from "../config.js";
 import { confirmAssessment } from "./assessmentStore.js";
 import { errorPayload } from "../errors.js";
+import { logger } from "../logging.js";
 loadConfig();
 const assessmentId = process.argv[2];
 if (!assessmentId || process.argv.length !== 3) {
-    console.error("Usage: patchwarden-confirm <full_assessment_id>");
-    console.error("The display-only assessment_short_id is not accepted.");
+    logger.info("Usage: patchwarden-confirm <full_assessment_id>");
+    logger.info("The display-only assessment_short_id is not accepted.");
     process.exit(1);
 }
 try {
-    const result = confirmAssessment(assessmentId);
+    const result = await confirmAssessment(assessmentId);
     console.log(JSON.stringify(result, null, 2));
 }
 catch (error) {
-    console.error(JSON.stringify(errorPayload(error), null, 2));
+    logger.error("confirm_assessment_failed", { error: errorPayload(error) });
     process.exit(1);
 }

package/dist/config.d.ts

@@ -17,6 +17,13 @@ export interface PatchWardenConfig {
     watcherStaleSeconds: number;
     toolProfile?: "full" | "chatgpt_core" | "chatgpt_direct" | "chatgpt_search";
     repoAliases?: Record<string, string>;
+    httpPort?: number;
+    http?: {
+        port?: number;
+        host?: string;
+        ownerTokenEnv?: string;
+    };
+    enableRunTaskTool?: boolean;
     enableAgentAssessment?: boolean;
     agentAssessmentTimeoutSeconds?: number;
     agentAssessmentMaxOutputBytes?: number;

package/dist/config.js

@@ -182,6 +182,40 @@ function normalizeConfig(config) {
         config.toolProfile !== "chatgpt_search") {
         throw new Error('toolProfile must be "full", "chatgpt_core", "chatgpt_direct", or "chatgpt_search"');
     }
+    if (config.repoAliases !== undefined) {
+        if (typeof config.repoAliases !== "object" || config.repoAliases === null || Array.isArray(config.repoAliases)) {
+            throw new Error("repoAliases must be an object mapping alias names to repository paths");
+        }
+        for (const [alias, target] of Object.entries(config.repoAliases)) {
+            if (typeof target !== "string") {
+                throw new Error(`repoAliases["${alias}"] must be a string`);
+            }
+        }
+    }
+    if (config.httpPort !== undefined) {
+        if (typeof config.httpPort !== "number" || !Number.isInteger(config.httpPort) || config.httpPort < 1 || config.httpPort > 65535) {
+            throw new Error("httpPort must be an integer from 1 to 65535");
+        }
+    }
+    if (config.http !== undefined) {
+        if (typeof config.http !== "object" || config.http === null || Array.isArray(config.http)) {
+            throw new Error("http must be an object");
+        }
+        if (config.http.port !== undefined) {
+            if (typeof config.http.port !== "number" || !Number.isInteger(config.http.port) || config.http.port < 1 || config.http.port > 65535) {
+                throw new Error("http.port must be an integer from 1 to 65535");
+            }
+        }
+        if (config.http.host !== undefined && typeof config.http.host !== "string") {
+            throw new Error("http.host must be a string");
+        }
+        if (config.http.ownerTokenEnv !== undefined && typeof config.http.ownerTokenEnv !== "string") {
+            throw new Error("http.ownerTokenEnv must be a string");
+        }
+    }
+    if (config.enableRunTaskTool !== undefined && typeof config.enableRunTaskTool !== "boolean") {
+        throw new Error("enableRunTaskTool must be a boolean");
+    }
     if (config.enableAgentAssessment !== undefined && typeof config.enableAgentAssessment !== "boolean") {
         throw new Error("enableAgentAssessment must be a boolean");
     }

package/dist/control/middleware/auth.d.ts

@@ -0,0 +1,10 @@
+/**
+ * Control Center auth middleware — control token validation.
+ *
+ * POST routes are gated behind an in-memory control token generated at boot.
+ * The token is exposed read-only via GET /control-token.json (handled by the
+ * server router) so the local dashboard can read it, but every mutating
+ * action must pass through `checkControlToken` first.
+ */
+import { type IncomingMessage } from "node:http";
+export declare function checkControlToken(req: IncomingMessage): boolean;

package/dist/control/middleware/auth.js

@@ -0,0 +1,8 @@
+import { controlToken } from "../shared.js";
+export function checkControlToken(req) {
+    const header = req.headers["x-patchwarden-control-token"];
+    const provided = Array.isArray(header) ? header[0] : header;
+    if (typeof provided !== "string" || provided.length === 0)
+        return false;
+    return provided === controlToken;
+}

package/dist/control/middleware/static.d.ts

@@ -0,0 +1,3 @@
+import { type ServerResponse } from "node:http";
+export declare function serveStatic(res: ServerResponse, urlPath: string): void;
+export declare function serveFavicon(res: ServerResponse): void;

package/dist/control/middleware/static.js

@@ -0,0 +1,65 @@
+/**
+ * Control Center static file middleware.
+ *
+ * Serves the bundled `ui/` directory (dashboard pages, partials, vendor
+ * assets, css) with path-traversal protection, plus the inline SVG favicon.
+ * Content types are mapped from `CONTENT_TYPES` in shared.ts.
+ */
+import { existsSync, readFileSync, statSync } from "node:fs";
+import { extname, isAbsolute, relative, resolve } from "node:path";
+import { CONTENT_TYPES, CONTROL_CENTER_FAVICON, errorMessage, sendJson, uiRoot, } from "../shared.js";
+export function serveStatic(res, urlPath) {
+    let candidate = "";
+    try {
+        const normalized = urlPath.startsWith("/") ? urlPath.slice(1) : urlPath;
+        let decoded;
+        try {
+            decoded = decodeURIComponent(normalized);
+        }
+        catch {
+            sendJson(res, 400, { error: "Invalid path encoding" });
+            return;
+        }
+        if (decoded.includes("\0")) {
+            sendJson(res, 400, { error: "Invalid path" });
+            return;
+        }
+        const segments = decoded.split("/").filter(Boolean);
+        if (segments.some((s) => s === "..")) {
+            sendJson(res, 403, { error: "Forbidden" });
+            return;
+        }
+        candidate = resolve(uiRoot, ...segments);
+        const rel = relative(uiRoot, candidate);
+        if (rel === "" || rel.startsWith("..") || isAbsolute(rel)) {
+            sendJson(res, 403, { error: "Forbidden" });
+            return;
+        }
+    }
+    catch (err) {
+        sendJson(res, 500, { error: errorMessage(err) });
+        return;
+    }
+    try {
+        if (!existsSync(candidate) || !statSync(candidate).isFile()) {
+            sendJson(res, 404, { error: "Not found" });
+            return;
+        }
+        const ext = extname(candidate).toLowerCase();
+        const contentType = CONTENT_TYPES[ext] || "application/octet-stream";
+        const content = readFileSync(candidate);
+        res.writeHead(200, { "Content-Type": contentType });
+        res.end(content);
+    }
+    catch (err) {
+        sendJson(res, 500, { error: errorMessage(err) });
+    }
+}
+export function serveFavicon(res) {
+    res.writeHead(200, {
+        "Content-Type": "image/svg+xml; charset=utf-8",
+        "Cache-Control": "public, max-age=86400",
+        "X-Content-Type-Options": "nosniff",
+    });
+    res.end(CONTROL_CENTER_FAVICON);
+}

package/dist/control/routes/audit.d.ts

@@ -0,0 +1,15 @@
+import { type ServerResponse } from "node:http";
+export type LogCategory = "core" | "direct" | "watcher" | "control-center";
+export declare function handleLogs(res: ServerResponse, category: LogCategory, tailLines: number): void;
+export declare function handleAudit(res: ServerResponse): void;
+/**
+ * Aggregate warnings from multiple sources:
+ *  - audit.json warnings/fail_checks/possible_false_positives/manual_verification_required
+ *  - stale task classification (heartbeat stale, collecting_artifacts stale, etc.)
+ *  - task status=failed_verification
+ *
+ * Each warning type is grouped into a bucket with affected task IDs. The
+ * handler is fault-tolerant: any per-task read failure is skipped, and a
+ * top-level failure returns an empty warnings array (never 500).
+ */
+export declare function handleWarnings(res: ServerResponse): void;