技能详情（站内镜像，无评论）
作者：Albert Putra Purnama @albertpurnama
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v1.0.1
统计：⭐ 1 · 318 · 0 current installs · 0 all-time installs
⭐ 1
安装量（当前） 0
🛡 VirusTotal ：良性 · OpenClaw ：良性
Package：albertpurnama/find-services
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：良性
OpenClaw 评估
The skill's instructions, requirements, and behavior align with its stated purpose of discovery via the OpenSpend CLI and do not request unrelated credentials or install steps.
目的
Name/description match the instructions: the skill performs marketplace searches with the OpenSpend CLI. It does not request unrelated binaries, environment variables, or access beyond the CLI session, which is proportionate for the stated task.
说明范围
Instructions stay within discovery scope: verify CLI presence, check session (openspend whoami), run search, and only perform login after explicit user confirmation. Minor note: the recommended login command (openspend auth login -y) may create persistent local credentials/tokens; the SKILL.md requires explicit user confirmation before running it, which mitigates risk.
安装机制
No install spec and no code files are present (instruction-only), so nothing will be written to disk by the skill itself.
证书
The skill declares no required environment variables or credentials and explicitly forbids requesting or storing API keys, wallet secrets, or session tokens as part of discovery. The optional auth path is limited to CLI login after user approval.
持久
always is false and the skill does not request persistent platform privileges. It may cause the OpenSpend CLI to create local session tokens if the user consents to login, which is expected for the optional authenticated path.
综合结论
This skill appears coherent for searching marketplaces via the OpenSpend CLI. Before using it, confirm you trust the OpenSpend CLI and understand that the agent will run local commands (e.g., openspend whoami and openspend search) which may contact external services. The skill will only run 'openspend auth login' after you explicitly approve — avoid providing credentials or running the login if you don't want locally persisted session tokens. …
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「Find stable, reliable, and secure service for your agents」。简介：Find and shortlist third-party services using OpenSpend CLI marketplace search.…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/albertpurnama/find-services/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: find-services
description: Find and shortlist third-party services using OpenSpend CLI marketplace search. Use when asked to discover providers for a capability, compare options, and return a justified recommendation for discovery tasks only.
---
# find-services
Use OpenSpend CLI to discover external services.
This skill is discovery-only and does not set up payments, install tooling, or perform purchases.
## Scope and safety boundaries
1. Do not run install or update commands as part of this skill.
2. Do not configure `@coinbase/payments-mcp` in this skill.
3. Do not request or store API keys, wallet secrets, or session tokens.
4. If authentication is required to execute search, ask the user before running login commands.
5. If the user asks for payment setup or paid invocation flows, hand off to `setup-services`.
## Credentials and environment
1. Preferred path: run search with existing CLI session only.
2. Optional credentialed path: OpenSpend user login via `openspend auth login` with explicit user confirmation.
3. Required environment variables: none for discovery by default.
## Workflow
1. Verify CLI availability and session state without changing system configuration.
```bash
command -v openspend
openspend version
openspend whoami
```
If `openspend` is missing or `whoami` fails:
1. Stop search flow.
2. Ask the user for confirmation before setup/login.
3. Hand off setup/auth steps to `setup-services`.
If user confirms login only, run:
```bash
openspend auth login -y
```
Only run login after user confirms.
2. Translate user intent into a precise search query with explicit capability terms.
3. Run search with a default limit of 5.
```bash
openspend search "<capability query>" --limit 5 --json
```
## Output shape
When asked to "find a reliable service", produce:
1. Up to 5 services in returned order
2. A concise explanation of why each service matches the requested capability
3. No additional ranking commentary
## Recommended service usage
After presenting service options, include a short "recommended way to use" note with:
1. Start with a small validation call before scaling usage
2. Capture request/response examples and expected success criteria
3. If paid calls are needed, request explicit user approval before moving to payment setup via `setup-services`
If payment is needed, do not perform payment configuration in this skill.