← 返回 Skills 市场
97
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install stock-analysis-7step
功能描述
使用7步分析法对中国A股进行全方位财务分析，自动生成包含ROE分解、盈利模式识别、行业对比和投资洞察的HTML报告。支持上海和深圳交易所所有股票。
安全使用建议
This package is an "encapsulated" client: it does not perform stock analysis locally but forwards your message to a remote Prana/Claw service which performs the analysis and returns results. Before installing or running it consider:
- Network exposure: the scripts will make network calls to the default base https://claw-uat.ebonex.io/ (or whatever NEXT_PUBLIC_URL you supply). Review and confirm the remote endpoint is trusted.
- Credential handling: if you don't provide PRANA_SKILL_PUBLIC_KEY/PRANA_SKILL_SECRET_KEY or PRANA_SKILL_API_KEY, the script will call GET /api/v1/api-keys and, by default, save the returned public_key:secret_key into config/api_key.txt on disk. If you do not want secrets written, set PRANA_SKILL_SKIP_WRITE_API_KEY=1 or provide keys via env vars.
- Data exfiltration: user input and frontmatter-derived parameters are sent to the remote service in POST /api/claw/agent-run; do not forward sensitive data you would not want external services to receive.
- Long-running behavior: the client may poll for results for a long time (defaults: 120s interval, up to 20 attempts, and HTTP timeouts up to 2 hours).
If you accept that the real analysis runs on the remote service and you trust that service and its base URL, the package is consistent with that design. If you expected a self-contained local analyzer or want to avoid storing secrets on disk or sending data externally, do not install/run this skill.
功能分析
Type: OpenClaw Skill
Name: stock-analysis-7step
Version: 1.0.0
The skill bundle is a legitimate 'thin client' designed to interface with the Prana/Claw remote API for A-share financial analysis. The Python and Node.js scripts (scripts/prana_skill_client.py and scripts/prana_skill_client.js) are responsible for managing API authentication, fetching credentials from a designated endpoint (claw-uat.ebonex.io), and forwarding user queries to the backend. The filesystem and network permissions requested in skill.yaml are consistent with the documented behavior of storing API keys in config/api_key.txt and communicating with the service provider. No evidence of malicious intent, obfuscation, or unauthorized data exfiltration was found.
能力评估
Purpose & Capability
The skill advertises a 7-step A-share financial analysis but the package contains no business logic — both Python/Node entry scripts are thin clients that forward user messages to a remote Prana/Claw service which performs the real work. This can be legitimate for an "encapsulated" skill, but it means the local package itself does not perform analysis; running it gives network access to the remote service.
Instruction Scope
The SKILL.md and scripts direct the agent to: read/write config/api_key.txt, automatically fetch API keys from a remote GET /api/v1/api-keys endpoint (DEFAULT_PRANA_BASE https://claw-uat.ebonex.io/), include x-api-key headers (public_key:secret_key) in subsequent requests, and POST user content to /api/claw/agent-run and /api/claw/agent-result. That means user messages and any context passed to the script are transmitted to the remote service; credentials may be auto-persisted to disk by default. The manifest declared no required env vars/credentials, but the runtime actually requires API keys (or will attempt to fetch them) — this mismatch is notable.
Install Mechanism
There is no custom install script; Node runner requires a single npm dependency (yaml) per package.json. This is a low-risk, standard dependency install (npm install). There are no downloads from arbitrary URLs or archive extraction in the package.
Credentials
Registry metadata lists no required env vars or primary credential, but the runtime supports/uses many env vars (PRANA_SKILL_PUBLIC_KEY, PRANA_SKILL_SECRET_KEY, PRANA_SKILL_API_KEY, NEXT_PUBLIC_URL, ACCOUNT_ID, etc.) and will automatically call GET /api/v1/api-keys to obtain a public_key/secret_key if not provided. The scripts will by default persist the fetched public_key:secret_key into config/api_key.txt. Requesting and storing these credentials is reasonable for a thin client, but the metadata omission (no declared credential) and the default write-to-disk behavior are proportionality concerns that users should be aware of.
Persistence & Privilege
The skill is not always: true. It does request network and filesystem permissions in skill.yaml and by default writes API credentials into config/api_key.txt. Writing its own credential file is expected for this wrapper, but it does create/modify local files and persists secrets unless PRANA_SKILL_SKIP_WRITE_API_KEY is set. Autonomous invocation is allowed (disable-model-invocation=false) which is standard; combined with network+credential behavior this increases blast radius if you enable the skill without understanding the remote endpoint.
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 stock-analysis-7step - 安装完成后，直接呼叫该 Skill 的名称或使用
/stock-analysis-7step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
v1.0.0
Initial release of A股财务分析 skill with 7-step analysis method.
- Provides comprehensive financial analysis for all stocks on Shanghai and Shenzhen exchanges.
- Automatically generates HTML reports including ROE decomposition, profit model identification, industry comparison, and investment insights.
- Supports both Python and Node.js clients for flexible integration.
- Includes documentation for authentication and accessing purchase records if the skill is paid.
元数据
常见问题
stock_analysis_7step 是什么？
使用7步分析法对中国A股进行全方位财务分析，自动生成包含ROE分解、盈利模式识别、行业对比和投资洞察的HTML报告。支持上海和深圳交易所所有股票。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 97 次。
如何安装 stock_analysis_7step？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install stock-analysis-7step」即可一键安装，无需额外配置。
stock_analysis_7step 是免费的吗？
是的，stock_analysis_7step 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
stock_analysis_7step 支持哪些平台？
stock_analysis_7step 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 stock_analysis_7step？
由 luokeer52（@luokeer52）开发并维护，当前版本 v1.0.0。
推荐 Skills