Skip to content

Self-hosted workflows for coding-agent delivery. Orchestrate CodeBuddy, Codex and Claude Code with scripts and human approval gates. Includes Local workflows for coding assistants.

docs: highlight local agent observability (#18)

Robert0622a3688dfd47ab
diff --git a/README.md b/README.md
index 7102249e..2dcbed76 100644
--- a/README.md
+++ b/README.md
@@ -115,6 +115,30 @@ workflow-summary.md
Resume an interrupted Classic workflow with `/resume-devflow {task_slug}` on CodeBuddy, Cursor, or Claude Code, and with `$devflow-codex resume {task_slug}` on Codex. Each installed host bundle documents its status, abort, and individual-stage commands.
+## See how agents work
+
+The Classic Edition for CodeBuddy includes a local observability dashboard. It turns the workflow evidence into a view of session activity, Token and cost trends, Skill/Rule usage, multi-agent dispatch, and Devflow stage attribution—useful for finding expensive runs, checking whether automation is actually being used, and reviewing how work moved between agents.
+
+![LoopForge local agent observability dashboard with demonstration data](docs/assets/agent-observability-dashboard.png)
+
+_Dashboard example using synthetic demonstration data._
+
+The dashboard is generated and served locally; it does not require an external telemetry service. Runtime logs can contain tool arguments and output excerpts, so LoopForge excludes them from Git by default and they should not be shared without review.
+
+<details>
+<summary>Open the local dashboard</summary>
+
+After running at least one CodeBuddy session from the project root:
+
+```bash
+python3 .codebuddy/skills/agent-observability/scripts/build_dashboard_data.py --project-root .
+python3 -m http.server 8765 --directory .codebuddy/skills/agent-observability/scripts/dashboard
+```
+
+Open `http://localhost:8765`. See the [observability quickstart](.codebuddy/skills/agent-observability/references/quickstart.md) for data sources, custom pricing, and troubleshooting.
+
+</details>
+
## When LoopForge is useful
LoopForge is especially useful when:
diff --git a/README.zh-CN.md b/README.zh-CN.md
index 8389420a..b70316ff 100644
--- a/README.zh-CN.md
+++ b/README.zh-CN.md
@@ -114,6 +114,30 @@ workflow-summary.md
Classic 工作流中断后,CodeBuddy、Cursor 和 Claude Code 使用 `/resume-devflow {task_slug}` 恢复,Codex 使用 `$devflow-codex resume {task_slug}`。状态查看、终止和单阶段命令见安装后对应宿主目录中的 README。
+## 看见 Agent 如何完成工作
+
+CodeBuddy 的 Classic Edition 内置本地可观测看板。它把落盘的流程证据变成可视化视图,集中呈现会话活动、Token 与成本趋势、Skill/Rule 使用情况、多 Agent 派发,以及 Devflow 阶段归因,方便定位高成本运行、确认自动化是否真正生效,并复盘任务如何在不同 Agent 之间流转。
+
+![使用演示数据的 LoopForge 本地 Agent 可观测看板](docs/assets/agent-observability-dashboard.png)
+
+_看板示例使用合成演示数据。_
+
+看板的数据聚合和页面服务都在本机完成,不依赖外部遥测服务。运行日志可能包含工具参数和输出片段,因此 LoopForge 默认将它们排除在 Git 之外;未经检查不应对外分享原始日志。
+
+<details>
+<summary>打开本地看板</summary>
+
+在项目根目录至少运行一次 CodeBuddy 会话后执行:
+
+```bash
+python3 .codebuddy/skills/agent-observability/scripts/build_dashboard_data.py --project-root .
+python3 -m http.server 8765 --directory .codebuddy/skills/agent-observability/scripts/dashboard
+```
+
+浏览器打开 `http://localhost:8765`。数据源、自定义模型定价和故障排查见[可观测性快速上手](.codebuddy/skills/agent-observability/references/quickstart.md)。
+
+</details>
+
## 什么时候适合使用
LoopForge 尤其适合这些任务:
diff --git a/docs/assets/agent-observability-dashboard.png b/docs/assets/agent-observability-dashboard.png
new file mode 100644
index 00000000..b6bb42c3
Binary files /dev/null and b/docs/assets/agent-observability-dashboard.png differ