Description

从185个AI角色中智能匹配最合适角色的工作流。当用户描述工作任务、需要找合适的AI角色来完成任务时触发。触发词：「角色推荐」、「智能匹配」、「用工作流」、「帮我找个角色」、「选个角色」。5步流程：1.用户描述任务 → 2.推荐2-3个角色含理由 → 3.用户确认选哪个 → 4.加载角色SOUL.md → 5.以...

README (SKILL.md)

Agency-Agent 智能角色推荐工作流

Name: Agency Role Match
Author: wzdongyifan-bit

核心原则

必须先推荐角色，等用户确认后才能开始工作。不能跳过推荐步骤直接输出成果。

5步工作流

第1步：用户描述任务

用户用自然语言描述想完成的工作，如：

"帮我分析亚马逊搜索词报表"
"写一个小红书推广文案"
"做一个竞品分析"

第2步：分析并推荐角色

从 ~/.openclaw/agency-agents/ 目录中搜索匹配的角色，推荐2-3个：

判断维度：

领域匹配：角色专长 vs 任务类型
技能匹配：角色技能栈 vs 任务需求
平台经验：是否有对应平台（Amazon/TikTok/小红书等）的实战

推荐格式（必须严格按此格式输出）：

针对「[任务描述]」任务，推荐以下角色：

🥇 [角色名] `角色ID`
   理由：为什么这个最合适

🥈 [角色名] `角色ID`
   理由：备选理由

🥉 [角色名] `角色ID`
   理由：第三个备选

你选哪个？
A. 用🥇
B. 换成🥈
C. 换成🥉
D. 两个都用

第3步：用户确认

等待用户选择（输入 A/B/C/D 或其他指示）

第4步：加载角色SOUL.md

根据用户选择，从 ~/.openclaw/agency-agents/[角色ID]/SOUL.md 读取角色设定

第5步：以角色身份工作

按照加载的角色SOUL.md，以该角色的身份、专业术语、沟通风格完成工作并输出成果

关键规则

推荐必须带理由：让用户知道为什么这个角色适合
推荐2-3个：太少没选择，太多选择困难
等用户确认：未经确认不能跳过第2步
加载SOUL后再工作：确保以正确的角色身份输出

角色目录

~/.openclaw/agency-agents/（185个角色，每个子目录含SOUL.md）

常用角色速查

任务类型	推荐角色ID
亚马逊搜索词分析	`paid-media-search-query-analyst`
增长策略	`marketing-growth-hacker`
亚马逊运营	`marketing-cross-border-ecommerce`
小红书运营	`marketing-xiaohongshu-specialist`
TikTok策略	`marketing-tiktok-strategist`
写文案	`marketing-content-creator`
竞品分析	`marketing-ecommerce-operator`
广告审计	`paid-media-auditor`
产品设计	`product-manager`

Usage Guidance

This skill will search and read files from ~/.openclaw/agency-agents/ and then adopt the behavior described in each role's SOUL.md. Before installing or enabling it: (1) Inspect the contents of ~/.openclaw/agency-agents/ and review any SOUL.md files for unexpected commands, external endpoints, or instructions that could leak data or trigger actions; (2) Ensure the directory and files are owned/trusted and have restrictive permissions; (3) If possible, run the skill in a sandbox or with limited agent permissions (no network/shell access) until you confirm the roles are safe; (4) Ask the skill author to declare required config paths (so the metadata matches runtime behavior) and to state limits on what SOUL.md can contain. If you cannot verify the role files' contents, treat the skill as higher risk.

Capability Analysis

Type: OpenClaw Skill Name: agency-role-match Version: 1.0.0 The skill bundle defines a standard workflow for an AI agent to recommend and load specialized sub-agent roles from a local directory (~/.openclaw/agency-agents/). The instructions in SKILL.md describe a logical 5-step process for task analysis and role selection without any evidence of malicious intent, data exfiltration, or unauthorized command execution.

Capability Assessment

ℹ Purpose & Capability

The name/description (match a role from a pool and act as that role) aligns with the runtime instructions which search ~/.openclaw/agency-agents/ and load SOUL.md files. However, the skill metadata declares no required config paths even though the SKILL.md explicitly requires access to a particular directory under the user's home. That mismatch is an incoherence.

⚠ Instruction Scope

The SKILL.md instructs the agent to read role files from ~/.openclaw/agency-agents/[roleID]/SOUL.md and then 'act as' that role. Reading arbitrary files under the user's home directory is exactly what the skill needs to work, but the instructions give the agent broad discretion to load and then adopt the behavior found in those files. If SOUL.md files contain instructions that trigger network calls, shell commands, or secrets exposure, the agent could perform actions beyond what the user expects. The SKILL.md does not declare or limit what the SOUL.md contents may contain.

✓ Install Mechanism

This is an instruction-only skill with no install spec and no external downloads. That minimizes disk-writing/install risk.

✓ Credentials

The skill declares no environment variables or credentials, and no sensitive external credentials are requested. This is proportionate to the stated purpose of selecting and loading local role files.

✓ Persistence & Privilege

always is false and the skill is user-invocable. It does not request permanent inclusion or system-wide configuration changes. The main privilege is file access to a local directory (which the SKILL.md relies on).

Version History

v1.0.0

- Initial release of agency-role-match workflow. - Matches user-described tasks to the most suitable AI roles from a pool of 185. - Recommends 2–3 roles with specific reasons and waits for user confirmation. - Loads the selected role’s SOUL.md settings before starting the task. - Outputs results strictly in the chosen role’s persona and expertise. - Enforces a 5-step process to ensure transparency and accurate role matching.

Metadata

Slug agency-role-match

Version 1.0.0

License MIT-0

All-time Installs 0

Active Installs 0

Total Versions 1

Frequently Asked Questions

What is Agency Role Match?

从185个AI角色中智能匹配最合适角色的工作流。当用户描述工作任务、需要找合适的AI角色来完成任务时触发。触发词：「角色推荐」、「智能匹配」、「用工作流」、「帮我找个角色」、「选个角色」。5步流程：1.用户描述任务 → 2.推荐2-3个角色含理由 → 3.用户确认选哪个 → 4.加载角色SOUL.md → 5.以... It is an AI Agent Skill for Claude Code / OpenClaw, with 136 downloads so far.

How do I install Agency Role Match?

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

Is Agency Role Match free?

Yes, Agency Role Match is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Agency Role Match support?

Agency Role Match is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Agency Role Match?

It is built and maintained by wzdongyifan-bit (@wzdongyifan-bit); the current version is v1.0.0.

More Skills

Agency Role Match