← 返回 Skills 市场
arch1904

OpenClaw Agent Creator

作者 arch1904 · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
1004
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install claw-agent-creator-archit
功能描述
Create new OpenClaw agents for Arch's multi-agent system. Use this skill when asked to create, add, or set up a new OpenClaw agent, or when adding an agent to the system defined in ~/.openclaw/. Covers the full lifecycle: directory creation, workspace files (SOUL.md, IDENTITY.md, etc.), openclaw.json config, Telegram routing (bindings + groups + mention patterns), cron job creation with proper prompt engineering, and gateway restart. Includes hard-won lessons from building the Wire (News) agent — the first non-default agent in the system. Also use when modifying existing agent configs, adding cron jobs to agents, or debugging agent routing issues.
使用说明 (SKILL.md)

OpenClaw Agent Creator

Create and configure agents for Arch's OpenClaw multi-agent system at ~/.openclaw/.

System Context

  • Owner: Archit (Arch), Linux user archit, timezone America/Denver
  • Gateway: Single process on port 18789 managing all agents
  • Bot: One Telegram bot shared across all agents — routing determines which agent handles which chat
  • Existing agents: Check ~/.openclaw/openclaw.jsonagents.list[] for current roster
  • Implementation history: See ~/.openclaw/implementation-docs/ for the Wire agent reference implementation

Agent Creation Workflow

1. Gather Requirements

Before creating anything, clarify with Arch:

  • Agent name and ID (lowercase, no spaces for ID)
  • Role and responsibilities (specific, not vague)
  • Model tier: cheap (Kimi K2.5 only) or full cascade (include Claude Sonnet)
  • Whether it needs a Telegram group for Q&A
  • Whether it needs cron jobs (what schedule, what tasks)
  • Whether heartbeat should be enabled or disabled

2. Stop the Gateway

openclaw gateway stop

MANDATORY before editing openclaw.json or cron/jobs.json. The gateway actively writes to jobs.json (updating job state after each cron run). Editing while the gateway runs causes race conditions and data loss.

3. Backup Config

cp ~/.openclaw/openclaw.json ~/.openclaw/openclaw.json.backup.$(date +%Y%m%d%H%M%S)

4. Create Directories

mkdir -p ~/.openclaw/workspace-\x3Cagent_id>/memory
mkdir -p ~/.openclaw/agents/\x3Cagent_id>/agent

NEVER reuse agentDir across agents — causes auth/session collisions.

5. Write Workspace Files

Use templates from assets/templates/ as starting points. Every agent needs:

File Purpose Required
SOUL.md Personality, role, responsibilities, behavioral modes Yes
IDENTITY.md Quick-reference card (name, role, emoji) Yes
USER.md About Arch (copy from any existing agent workspace) Yes
AGENTS.md Workspace rules (boot sequence, memory, safety) Yes
HEARTBEAT.md Periodic task checklist (or comment if disabled) Yes

SOUL.md is the most important file. Be specific about responsibilities. Include behavioral modes if the agent operates differently in different contexts (e.g., briefing mode vs chat mode).

6. Edit openclaw.json — Agent Entry

Add to agents.list[]. See references/config-schema.md for all valid fields.

Minimal entry:

{
  "id": "\x3Cagent_id>",
  "name": "\x3CDisplay Name>",
  "workspace": "/home/archit/.openclaw/workspace-\x3Cagent_id>",
  "agentDir": "/home/archit/.openclaw/agents/\x3Cagent_id>/agent",
  "identity": { "name": "\x3CDisplay Name>" }
}

Common additions:

  • "model" — Override the default model cascade. Exclude expensive models for worker agents.
  • "heartbeat": { "every": "0" } — Disable heartbeat for cron-only agents.
  • "groupChat": { "mentionPatterns": ["@\x3Cid>", "@\x3CName>"] } — Enable @mentions in groups.

Only ONE agent should have "default": true (currently Fossil). The default agent receives all unrouted messages.

7. Edit openclaw.json — Telegram Routing (if needed)

THREE separate config changes are required. Missing any one causes silent failures. See references/telegram-routing.md for the full explanation.

  1. Group config in channels.telegram.groups:

    "-100XXXXXXXXXX": { "requireMention": false }
    
  2. Binding in bindings[]:

    { "agentId": "\x3Cid>", "match": { "channel": "telegram", "peer": { "kind": "group", "id": "-100XXXXXXXXXX" } } }
    
  3. Mention patterns on the agent entry (already done in step 6 if groupChat was added).

8. Create Cron Jobs (if needed)

Edit cron/jobs.json. Every cron job prompt MUST include:

  • Dynamic group ID resolution preamble (NEVER hardcode Telegram group IDs):
    FIRST: Resolve your Telegram group ID by running:
    jq -r '.bindings[] | select(.agentId == "\x3Cagent_id>") | .match.peer.id' ~/.openclaw/openclaw.json
    Use the output as the target for all Telegram messages in this task.
    
  • Date injection: $(date '+%A, %B %d, %Y') after the preamble
  • Explicit constraints: source allowlists, recency rules, format templates
  • Delivery instructions: use target='\x3CAGENT_GROUP_ID>' placeholder (resolved by the preamble)

This self-healing pattern ensures cron jobs survive Telegram group ID migrations. See references/prompt-patterns.md for full patterns and references/telegram-routing.md for why this matters.

Critical: If copying files or prompts from another agent's workspace, grep for hardcoded paths and update them.

9. Restart Gateway and Verify

openclaw gateway start

Verify in logs:

  • Agent registered: agent registered: \x3Cid>
  • Messages route correctly: lane enqueue: lane=session:agent:\x3Cid>:...

If messages to a Telegram group show skip: no-mention, the channels.telegram.groups config is missing (see references/bugs-and-pitfalls.md).

Reference Files

File When to Read
references/config-schema.md When writing agent config or cron jobs
references/telegram-routing.md When setting up Telegram group routing
references/prompt-patterns.md When writing cron job prompts
references/bugs-and-pitfalls.md When debugging issues or before any config edit

Template Files

Starter templates for workspace files are in assets/templates/. Copy and customize per agent.

安全使用建议
This skill appears to do what it says (create and configure OpenClaw agents), but review a few things before installing or running it: - Verify tool availability: the instructions assume the openclaw CLI and jq (and a POSIX shell) are present. The manifest lists no required binaries — make sure those tools exist on your system or the scripts/prompts will fail. - Review and limit autonomy: AGENTS.md contains the phrase 'Don't ask permission. Just do it.' If you don't want the agent making config changes or restarting the gateway without confirmation, require human approval before executing any stop/start, file edits, or cron creation steps. - Inspect workspace files for secrets: the skill instructs agents to read memory and workspace files (memory/YYYY-MM-DD.md, MEMORY.md). Those could contain sensitive information. Clean or audit those files before letting an automated flow read them, or add explicit rules to avoid reading secrets. - Back up before changing config: the skill advises backups; follow that. Ensure you have a good backup and/or test in a non-production environment first. - Cron prompts and shell substitution: cron prompt patterns include shell substitutions and instructions to run jq at runtime. Understand how those prompts will be stored and executed, and confirm they can't be abused to run unintended commands. If you accept those caveats and confirm the required CLI tools are present, the skill is reasonably coherent for its purpose. If you are unsure about autonomous edits or exposing local memory files, treat the skill as potentially risky and restrict its use to manual, guided runs only.
功能分析
Type: OpenClaw Skill Name: claw-agent-creator-archit Version: 1.0.0 The skill bundle is designed for OpenClaw agent system management, which inherently requires privileged operations like file system access and process control. It explicitly instructs the AI agent to execute shell commands (e.g., `openclaw gateway stop`, `cp`, `mkdir`, `jq`, `date`) as part of its workflow, as detailed in `SKILL.md` and `references/prompt-patterns.md`. Specifically, `references/prompt-patterns.md` shows a 'Report Prompt' pattern that instructs the agent to 'Run: 1) [cmd1] 2) [cmd2] ...', indicating the agent is designed to execute arbitrary commands provided in prompts. This direct instruction for shell command execution, especially with dynamic inputs and arbitrary commands, creates a critical prompt injection vulnerability, allowing for potential Remote Code Execution if a malicious user can inject commands into the agent's prompts.
能力评估
Purpose & Capability
The skill's name/description (create and configure OpenClaw agents in ~/.openclaw/) aligns with the actions described in SKILL.md (create dirs, edit openclaw.json, create cron jobs, restart gateway). However, the manifest lists no required binaries or tools even though the instructions assume availability of the openclaw CLI and utilities like jq and standard shell tools (cp, mkdir). This is an implementation omission that can cause failures or hide additional implicit requirements.
Instruction Scope
SKILL.md instructs the agent to stop/start the gateway, edit ~/.openclaw/openclaw.json, create workspace directories and files, and author cron job prompts that run shell substitutions and use jq to resolve Telegram group IDs. These operations are within the skill's scope, but the docs also tell agents (in AGENTS.md) 'Don't ask permission. Just do it.' and instruct agents to read workspace memory files (memory/YYYY-MM-DD.md, MEMORY.md). That encourages autonomous modification of system config and reading potentially sensitive local data. The combination — automated edits to config + reading workspace memory files that may contain secrets + explicit advice to act without asking — is a notable scope creep/risk for users who expect manual confirmation.
Install Mechanism
This is an instruction-only skill with no install spec and no code files, so nothing is written to disk by the skill itself. That's the lowest-risk install mechanism. The only concern is the mismatch between the instructions' implicit tool requirements (openclaw, jq, shell utilities) and the manifest's 'required binaries: none.'
Credentials
The skill requests no environment variables, no credentials, and no config paths in its manifest. The runtime instructions operate on user-local OpenClaw files under ~/.openclaw/, which is proportional to the stated purpose. No unrelated external credentials or remote endpoints are requested.
Persistence & Privilege
The skill is not marked always:true and does not request persistent system-wide privileges. It does, however, instruct the agent to perform privileged-seeming actions within the user's OpenClaw installation (editing openclaw.json, stopping/starting the gateway, creating cron jobs). Because the skill's text encourages 'just do it' behavior, consider the risk if the skill is invoked autonomously by the agent without interactive confirmation.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install claw-agent-creator-archit
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /claw-agent-creator-archit 触发
  4. 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
v1.0.0
- Initial release of the OpenClaw Agent Creator for Arch's multi-agent system. - Supports full agent lifecycle: directory setup, workspace file creation (SOUL.md, IDENTITY.md, etc.), and openclaw.json configuration. - Guides proper Telegram routing: group config, bindings, and mention pattern settings. - Includes instructions for creating robust cron jobs with prompt patterns that handle Telegram group ID changes. - Embeds best practices from building the Wire (News) agent and advice on debugging routing issues.
元数据
Slug claw-agent-creator-archit
版本 1.0.0
许可证
累计安装 1
当前安装数 1
历史版本数 1
常见问题

OpenClaw Agent Creator 是什么?

Create new OpenClaw agents for Arch's multi-agent system. Use this skill when asked to create, add, or set up a new OpenClaw agent, or when adding an agent to the system defined in ~/.openclaw/. Covers the full lifecycle: directory creation, workspace files (SOUL.md, IDENTITY.md, etc.), openclaw.json config, Telegram routing (bindings + groups + mention patterns), cron job creation with proper prompt engineering, and gateway restart. Includes hard-won lessons from building the Wire (News) agent — the first non-default agent in the system. Also use when modifying existing agent configs, adding cron jobs to agents, or debugging agent routing issues. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1004 次。

如何安装 OpenClaw Agent Creator?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install claw-agent-creator-archit」即可一键安装,无需额外配置。

OpenClaw Agent Creator 是免费的吗?

是的,OpenClaw Agent Creator 完全免费(开源免费),可自由下载、安装和使用。

OpenClaw Agent Creator 支持哪些平台?

OpenClaw Agent Creator 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 OpenClaw Agent Creator?

由 arch1904(@arch1904)开发并维护,当前版本 v1.0.0。

💬 留言讨论