← 返回 Skills 市场
hackstoic

HK IPO Decision Engine

作者 hackstoic · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ✓ 安全检测通过
75
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install hkipo-decision-engine
功能描述
Score Hong Kong IPOs, build decision cards, and batch-evaluate symbols or watchlists with explainable factors and execution guidance. Use when the user wants...
使用说明 (SKILL.md)

HK IPO Decision Engine

Use this skill for action-oriented decisions.

Runtime

This publish bundle includes the required CLI runtime under runtime/hkipo-next.

From the skill folder:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next ...

By default these commands persist review records in ~/.hkipo-next/data/hkipo.db.

Preconditions

  • Prefer to have an active parameter version.
  • Prefer to have a usable profile or explicit CLI overrides for risk, budget, and financing preference.
  • If the user only wants rough market discovery, do not jump straight to this skill.

Choose the Right Command

  • Use score for factor-level explanation.
  • Use decision-card for a concise execution plan.
  • Use batch for multiple symbols or the saved watchlist.

Commands

Score one symbol:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next score 2476 --format json

Score with explicit scenario overrides:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next score 2476 \
  --parameter-version v0003 \
  --risk-profile aggressive \
  --max-budget-hkd 120000 \
  --financing-preference margin \
  --format json

Generate one decision card:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next decision-card 2476 --format markdown

Batch explicit symbols:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next batch 2476 2613 2590 --format json

Batch the current watchlist:

cd \x3Cskill_dir>
uv run --directory runtime/hkipo-next hkipo-next batch --watchlist --format json

Output Cues

  • score returns factor contributions, cost breakdown, assumptions, and risk disclosure.
  • decision-card adds a headline, position suggestion, and exit plan.
  • partial outputs are often still usable; they usually mean source conflicts or incomplete upstream data, not total failure.

Companion Skills

  • Use $hkipo-parameter-manager if there is no active parameter version or the user wants tuning.
  • Use $hkipo-review-optimizer after outcomes are known and the user wants to refine the model.
安全使用建议
This package appears internally consistent with its stated purpose, but review these points before installing: - The skill runs a bundled CLI (hkipo-next) via the 'uv' runner and will perform network calls to many public data sources (e.g., aipo.myiqdii.com, aastocks.com, hkex, sinajs/qt.gtimg endpoints). Expect outbound HTTP(s) traffic when using it. - It reads and writes files under your home directory (~/.hkipo-next/config/*.json and ~/.hkipo-next/data/hkipo.db). Those files will contain watchlists, profiles, and persisted review history — inspect their contents if you care about local privacy. - The installer uses Homebrew to install 'uv'. If you have concerns, verify what the 'uv' formula points to in your Homebrew tap (source/repository) before installing. - If you need stronger isolation, run the skill inside a constrained environment (container or VM) or inspect the included source code (present in runtime/hkipo-next) — the bundle is large but human-readable and uses standard libraries (httpx, BeautifulSoup) not obfuscated loaders. - If anything would change this assessment: evidence that the 'uv' formula downloads arbitrary/external code during install, or that the runtime contains hidden endpoints / credential exfiltration logic. If you want, I can list the network endpoints called by the code and the paths that will be created/read under ~/.hkipo-next.
功能分析
Type: OpenClaw Skill Name: hkipo-decision-engine Version: 0.1.0 The hkipo-decision-engine skill is a comprehensive tool for analyzing Hong Kong IPOs, providing scoring, decision cards, and batch evaluation. The codebase is well-structured, utilizing Pydantic for data validation, SQLite for persistence, and httpx/requests for fetching data from legitimate financial sources (HKEX, AAStocks, Futu, and AiPO). It includes a sophisticated caching layer with file locking and atomic writes. No evidence of data exfiltration, malicious execution, or prompt injection was found; the 'import suggestions' feature is a legitimate mechanism for model refinement based on historical performance. The code follows professional development standards and aligns perfectly with its stated purpose.
能力评估
Purpose & Capability
Name/description align with the actual contents: the bundle includes scoring_service, decision_service, renderers, and many adapters (aipo, aastocks, hkex, jisilu, futu, tradesmart, etc.). Requiring the 'uv' runtime and local config/db files (~/.hkipo-next/...) is coherent with a packaged CLI that persists review records.
Instruction Scope
SKILL.md instructs the agent to run the included CLI via 'uv run --directory runtime/hkipo-next hkipo-next ...'. The runtime will access network data from multiple public endpoints (e.g., aipo.myiqdii.com, aastocks.com, hkex, suggest3.sinajs.cn, qt.gtimg.cn) and persists data in ~/.hkipo-next/data/hkipo.db and reads ~/.hkipo-next/config/*.json. This behavior is expected for the skill but means the agent will perform broad network scraping and local file reads/writes specific to this tool.
Install Mechanism
Install spec is a Homebrew formula for 'uv' (the CLI runner referenced throughout SKILL.md). Homebrew is a standard distribution mechanism; no arbitrary URLs or archive extraction are used. Installing 'uv' via brew is proportionate to the stated runtime requirement.
Credentials
The skill does not request environment variables or external credentials, which is consistent. It does require/configure local config and a sqlite DB under ~/.hkipo-next — these files will be created/used by the CLI and may contain user preferences, watchlists, and review history. The skill will therefore read/write user home files but does not request unrelated credentials.
Persistence & Privilege
always:false and no elevated privileges are requested. The skill installs a runtime binary (uv) and runs a local CLI, and it persists data under its own ~/.hkipo-next directory; it does not declare modifications to other skills or system-wide configurations.
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 hkipo-decision-engine
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /hkipo-decision-engine 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
v0.1.0
Initial ClawHub release
元数据
Slug hkipo-decision-engine
版本 0.1.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

HK IPO Decision Engine 是什么?

Score Hong Kong IPOs, build decision cards, and batch-evaluate symbols or watchlists with explainable factors and execution guidance. Use when the user wants... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 75 次。

如何安装 HK IPO Decision Engine?

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

HK IPO Decision Engine 是免费的吗?

是的,HK IPO Decision Engine 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

HK IPO Decision Engine 支持哪些平台?

HK IPO Decision Engine 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 HK IPO Decision Engine?

由 hackstoic(@hackstoic)开发并维护,当前版本 v0.1.0。

💬 留言讨论