← Back to Skills Marketplace
edoserbia

Agent Autopilot

by edoserbia · GitHub ↗ · v1.4.1
cross-platform ⚠ suspicious
5783
Downloads
1
Stars
72
Active Installs
6
Versions
Install in OpenClaw
/install agent-autopilot
Description
Self-driving agent workflow with heartbeat-driven task execution, day/night progress reports, and long-term memory consolidation. Integrates with todo-manage...
Usage Guidance
Install only in a dedicated, bounded workspace. Define clear stop conditions and require review before MEMORY.md cleanup, git commits, external tool use, or high-impact file changes. Verify the copied todo-management skill source before enabling heartbeat automation.
Capability Analysis
Type: OpenClaw Skill Name: agent-autopilot Version: 1.4.1 The skill bundle instructs the AI agent to execute shell commands, specifically `bash {todoBaseDir}/scripts/todo.sh ...`. While this is necessary for its stated purpose of task management, the template `bash {todoBaseDir}/scripts/todo.sh entry create "..." --group="{项目名}"` in `SKILL.md` and `HEARTBEAT.md` presents a shell injection vulnerability. If the AI agent, when autonomously creating new tasks, generates a `{项目名}` value from untrusted input or its own analysis that contains shell metacharacters, this could lead to arbitrary command execution. This is a significant risk, even if not intentionally malicious by the skill developer. The skill also grants the agent broad file system and `git` access, which, while necessary for its function, increases the attack surface if the agent is compromised or misdirected.
Capability Assessment
Purpose & Capability
The purpose is coherent: heartbeat-driven task execution, progress reporting, and memory consolidation. The concern is that the core capability is an open-ended autonomous loop that must keep progressing until goals are reached.
Instruction Scope
The runtime instructions tell the agent not to wait for direction, to create new tasks when goals are unmet, and to continue on every heartbeat, but they provide limited stop criteria, approval gates, or containment for what work may be performed.
Install Mechanism
Installation is a visible user-run shell script that creates workspace skills and memory directories, initializes memory/report-state.json, and copies todo-management from a local or global install. It is not obfuscated, but the copied dependency is not pinned or verified.
Credentials
Workspace todo writes, daily logs, reports, and MEMORY.md edits are purpose-aligned, but examples also encourage automatic git commits, external tool delegation, and ongoing project changes without clear review boundaries.
Persistence & Privilege
The skill creates durable operational state and instructs periodic updates to report-state.json and MEMORY.md, including cleanup of old memory content, which can steer future agent behavior across sessions.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install agent-autopilot
  3. After installation, invoke the skill by name or use /agent-autopilot
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.4.1
缩短description,修复ClawHub页面渲染溢出问题
v1.4.0
三个定时任务全部补充完整驱动性提示词:①自驱动循环有明确行为指令(立刻工作/不要HEARTBEAT_OK/自己判断) ②汇报有完整判断逻辑(白天/夜间/首报/周期/总结/即时) ③记忆维护有完整流程指令(提炼/清理/更新);HEARTBEAT模板从简单清单升级为可直接复制使用的完整提示词
v1.3.0
重构为3个明确的定时任务:①自驱动执行循环(每次heartbeat~30min) ②进度汇报(白天每N小时/夜间静默) ③长期记忆维护(每6小时);简化report-state.json为扁平结构;统一HEARTBEAT.md模板格式
v1.2.0
重构汇报机制:固定时间窗口改为白天/夜间间隔模式,白天每N小时汇报+夜间静默,report-state.json简化,去除茄子/老大等硬编码,所有模板通用化,新增多种汇报格式(首报/周期/总结/紧急),记忆维护与汇报周期联动
v1.1.0
新增定时进度汇报(晨报/午报/晚报/夜报)、长期记忆维护(每日/每3天/每周)、report-state.json 状态跟踪、完整 HEARTBEAT.md 模板、init.sh 自动初始化汇报状态
v1.0.0
初始发布:自驱动agent工作流,集成heartbeat+todo+汇报机制
Metadata
Slug agent-autopilot
Version 1.4.1
License
All-time Installs 198
Active Installs 72
Total Versions 6
Frequently Asked Questions

What is Agent Autopilot?

Self-driving agent workflow with heartbeat-driven task execution, day/night progress reports, and long-term memory consolidation. Integrates with todo-manage... It is an AI Agent Skill for Claude Code / OpenClaw, with 5783 downloads so far.

How do I install Agent Autopilot?

Run "/install agent-autopilot" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Agent Autopilot free?

Yes, Agent Autopilot is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Agent Autopilot support?

Agent Autopilot is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Agent Autopilot?

It is built and maintained by edoserbia (@edoserbia); the current version is v1.4.1.

💬 Comments