← 返回 Skills 市场
587
总下载
1
收藏
4
当前安装
10
版本数
在 OpenClaw 中安装
/install task-ledger
功能描述
Durable workflow layer for OpenClaw long-running work. Use when tasks are multi-stage, recoverable, parallelized across sub-agents/ACP, use background exec o...
安全使用建议
This toolkit appears coherent and local-only: it creates and updates task JSON files and log/output directories under your workspace and uses python3/bash. Before installing or running it, review the bundled scripts and templates to confirm the file paths are acceptable for your workspace and that you are happy with local disk writes to tasks/, logs/, and outputs/. The skill does not request secrets or perform network downloads, but standard caution applies: do not install into a workspace containing sensitive production files without backing them up, and confirm any task actions that perform external side effects (deploy/restart) before executing.
功能分析
Type: OpenClaw Skill
Name: task-ledger
Version: 0.3.2
The task-ledger bundle is a comprehensive and legitimate toolkit designed to provide a durable workflow management layer for OpenClaw agents. It consists of shell and Python scripts (e.g., new-task.sh, update-task.py, task-doctor.py) that manage task states, dependencies, and execution logs via local JSON files. The code is well-structured, lacks any network-reaching calls or obfuscation, and focuses entirely on its stated purpose of enabling recoverable, multi-stage agent workflows. No indicators of malicious intent, data exfiltration, or harmful prompt injection were found.
能力评估
Purpose & Capability
Name/description match the actual artifacts: a filesystem-backed toolkit for durable task objects. Required binaries (python3, bash) are appropriate for the included scripts. No unrelated credentials, binaries, or config paths are requested.
Instruction Scope
SKILL.md stays within scope: it instructs copying bundled toolkit files into the workspace, creating tasks/logs/outputs directories, and using the provided scripts to manage task JSONs. The scripts only read/write local workspace files and do not attempt to read unrelated system config or environment secrets. The instructions caution not to overwrite user-modified files and to confirm plans with the user.
Install Mechanism
There is no external install-from-URL. An included install.sh copies bundled files into the workspace only if missing. No external downloads, package manager installs, or archive extraction are performed.
Credentials
The skill requires no environment variables or credentials. The scripts operate on workspace directories (tasks/, logs/, outputs/) and only store/modify JSON task files and logs; they do not access or exfiltrate secrets, network endpoints, or unrelated system credentials.
Persistence & Privilege
always is false and the skill is user-invocable. The toolkit writes files only under the workspace (tasks/, logs/, outputs/, scripts/, task-templates/) and does not modify other skills or system-wide agent settings.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install task-ledger - 安装完成后，直接呼叫该 Skill 的名称或使用
/task-ledger触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v0.3.2
Add minimum task-level dependency and readiness semantics with explicit readiness config, upgraded task-ready/task-start-if-ready behavior, and support for all-succeeded, wait-all, fail-fast, and partial-allowed workflows.
v0.3.1
Add minimum parent-task summary aggregation in task-export.py, including child status counts, readiness signals, and per-child summaries in short and markdown exports without changing the task schema.
v0.3.0
Reposition Task Ledger for the OpenClaw 3.7 era as a durable workflow layer; add explicit usage boundaries, parent/child parallel orchestration guidance, and clearer relationships to ACP, sub-agents, exec, and cron.
v0.2.4
Fix final-stage advance so task-advance.py clears stage instead of producing a redundant self-transition; preserve explicit close guidance after the last stage completes.
v0.2.3
Add short-first reporting support with reporting schema fields, chat-friendly short export, and file-backed long reports written to outputs/<taskId>/report.md. Sync the published skill package to include reporting-aware workflow support.
v0.2.2
Add readiness-aware scheduling with task-ready and task-start-if-ready, plus relation/dependency graph export in markdown, dot, and json. Strengthen doctor scheduling diagnostics and sync the latest local toolkit into the published package.
v0.2.1
Add task relations and tree view, stronger dependency-aware doctor checks, human-friendly task export, Markdown resume summaries, improved process binding semantics, and legacy task migration helper.
v0.2.0
Add richer task schema with lifecycle, blocking, rollback, and notification fields. Add recovery verification and resume summary helpers. Add subtask/cron binding helpers. Add operational templates for safe gateway restart and service deployment.
v0.1.1
Improve discoverability: clearer long-running task, checkpoint, recovery, and resumable workflow wording; add better search tags.
v0.1.0
Initial release: persistent checkpointed long-task workflow for OpenClaw with task skeletons, stage tracking, recovery helpers, and bootstrap installer.
元数据
常见问题
Task Ledger 是什么？
Durable workflow layer for OpenClaw long-running work. Use when tasks are multi-stage, recoverable, parallelized across sub-agents/ACP, use background exec o... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 587 次。
如何安装 Task Ledger？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install task-ledger」即可一键安装，无需额外配置。
Task Ledger 是免费的吗？
是的，Task Ledger 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Task Ledger 支持哪些平台？
Task Ledger 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Task Ledger？
由 Leon Ge（@gejiliang）开发并维护，当前版本 v0.3.2。
推荐 Skills