← 返回 Skills 市场
139
总下载
0
收藏
0
当前安装
6
版本数
在 OpenClaw 中安装
/install openclaw-laravel-creem-agent-skill
功能描述
Creem payment store assistant — query subscriptions, customers, transactions, products, run heartbeat checks, and manage a payment store via a local Laravel...
安全使用建议
This skill is coherent with its purpose but can perform real payment-store actions by forwarding user text to a local HTTP endpoint. Before installing: 1) Verify that the Laravel Creem Agent service at 127.0.0.1:8000 is the trusted local service you expect and is not exposed to external networks. 2) Ensure the endpoint requires appropriate authentication/ACLs and that the agent host is secured. 3) Consider disabling autonomous invocation for this skill or require explicit confirmation for destructive intents (cancel/create) in OpenClaw to prevent accidental changes. 4) Test using read-only queries first and enable audit/logging on the endpoint so you can review actions the agent performed. 5) If you use channels (Telegram) follow OpenClaw docs to restrict who can send commands. If you want, provide the endpoint's access controls or a read-only mode and I can suggest concrete guardrails.
功能分析
Type: OpenClaw Skill
Name: openclaw-laravel-creem-agent-skill
Version: 1.0.5
The skill acts as a functional bridge between the OpenClaw agent and a local Laravel-based payment management service (Creem Agent). It uses curl to interact with a local HTTP endpoint (127.0.0.1:8000) to query store data like transactions and subscriptions. The instructions in SKILL.md are designed to ensure the agent prioritizes the tool over its internal knowledge and correctly handles JSON escaping, with no evidence of data exfiltration, malicious execution, or unauthorized system access.
能力评估
Purpose & Capability
The name/description match the behavior in SKILL.md: all actions are routed to a hard‑wired local Laravel endpoint. The only required binary is curl, which is appropriate for the described HTTP forwarding behavior.
Instruction Scope
SKILL.md strictly instructs the agent to POST user text to http://127.0.0.1:8000/creem-agent/chat and relay the endpoint's `response` verbatim. It does not read files or extra env vars. Note: the endpoint supports management intents (cancel_subscription, create_checkout, switch_store, etc.), so forwarded user text can result in state‑changing or destructive operations if not constrained or confirmed.
Install Mechanism
Instruction-only skill with no install spec and no code files. This is low-risk from an install perspective — nothing is downloaded or written by the skill itself.
Credentials
The skill declares no required environment variables, credentials, or config paths. README mentions general OpenClaw/Telegram configuration but those are not required by the skill itself and are typical for OpenClaw deployments.
Persistence & Privilege
always:false (not force-included). The platform default allows autonomous invocation, so the agent could call this skill without user confirmation — combined with the endpoint's ability to perform management actions, this raises an operational-risk consideration (accidental/cascading changes) but it is not an incoherence in the skill design.
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 openclaw-laravel-creem-agent-skill - 安装完成后，直接呼叫该 Skill 的名称或使用
/openclaw-laravel-creem-agent-skill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
v1.0.5
- Clarified intent preservation: Added explicit rules to maintain measurement, time, status, and ranking phrases from user questions.
- Expanded routing heuristics: Detailed new routing table and fallback rules to ensure precise query forwarding and accurate aggregation.
- Improved count handling: Specified that when a list is returned for a count/quantity question, count the items and reply with the number, not the raw list.
- Added and strengthened anti-patterns: Explicitly prohibit relaying transaction lists when a count is asked for, and clarified other disallowed behaviors.
- Reorganized and streamlined documentation for easier reference and improved precision during multi-step query handling.
v1.0.4
openclaw-laravel-creem-agent-skill 1.0.4 changelog
- Clarified that all store, payment, and subscription queries must be sent to the Laravel endpoint first.
- Added rules prohibiting fallback to local data, workspace files, saved notes, or database exports before calling the Laravel endpoint.
- Explicitly disallow generic disclaimers about missing local context or data when answering store-related questions.
- Enhanced anti-pattern and precision rules to prevent asking users about transaction storage or data sources before calling the endpoint.
- Updated multi-step query examples and rules to reinforce endpoint-first behavior.
v1.0.3
- Updated instructions to emphasize sending the user's exact question to the endpoint without simplification, especially for filtered/count questions (e.g., time periods, success/failure status).
- Added a new "Routing heuristics for low-cost models" section with detailed examples and anti-patterns to avoid over-generalizing user queries.
- Clarified that for multi-step queries, the exact question should be forwarded first, and only fallback to intent decomposition if the endpoint cannot answer directly.
- Introduced new "Precision rules" requiring important qualifiers (like "successful", "today", "top") to be preserved when forwarding.
- Strengthened prohibition against answering or replacing user queries with assistant reasoning when the endpoint can answer them.
- No file or implementation changes—documentation update improving query fidelity and robustness.
v1.0.2
Version 1.0.2
- Documentation streamlined: Removed operator override instructions (`AGENT_URL`), making endpoint, curl examples, and setup simpler.
- Quick self-test and usage examples now use the hardcoded local endpoint URL for clarity.
- No changes to source code or functional behavior; update is limited to documentation cleanup.
v1.0.1
- Default endpoint is now `http://127.0.0.1:8000/creem-agent/chat`; environment variable `AGENT_URL` is optional.
- Updated curl command examples to fall back to the default endpoint if `AGENT_URL` is not set.
- Environment variable requirement for `AGENT_URL` removed from metadata; only `curl` is now required.
- Updated documentation for endpoint configuration and connection instructions.
- No code or logic changes—documentation and metadata only.
v1.0.0
openclaw-laravel-creem-agent-skill 1.0.0 – Initial Release
- Introduces an assistant to manage and query Laravel Creem payment stores via a local HTTP endpoint.
- Supports subscription, customer, transaction, product queries, and store monitoring (heartbeat).
- Handles both single-intent and multi-step workflows through chained endpoint calls.
- Ensures all store/payment questions are answered strictly via the configured HTTP endpoint.
- Includes setup requirements, usage guidance, example curl calls, and best-practice rules for error handling and multi-store scenarios.
元数据
常见问题
Laravel Creem Agent 是什么？
Creem payment store assistant — query subscriptions, customers, transactions, products, run heartbeat checks, and manage a payment store via a local Laravel... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 139 次。
如何安装 Laravel Creem Agent？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install openclaw-laravel-creem-agent-skill」即可一键安装，无需额外配置。
Laravel Creem Agent 是免费的吗？
是的，Laravel Creem Agent 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Laravel Creem Agent 支持哪些平台？
Laravel Creem Agent 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（linux, darwin）。
谁开发了 Laravel Creem Agent？
由 romansh（@romansh）开发并维护，当前版本 v1.0.5。
推荐 Skills