← Back to list
alphaear-signal-tracker Checked
Track finance investment signal evolution and update logic based on new finance market informatio…
5.6Overall rating
Installation
🤖 Install via AI
Copy the prompt below and send it to your AI assistant (e.g. Claude Code) — it will follow the instructions and install automatically.
Install the "alphaear-signal-tracker" skill by following the instructions at https://skill123.me/install/alphaear-signal-tracker.
⌨️ Command line install
Run in your terminal — downloads and installs to ~/.claude/skills/.
curl -fsSL https://skill123.me/install/alphaear-signal-tracker.sh | bash
📦 Download ZIP
Download the zip and extract it into your skills directory (e.g. ~/.claude/skills/), then restart your session.
⬇ Download v1.0.0 · 101 KBAbout this skill
Overview
Track finance investment signal evolution and update logic based on new finance market information. Use when monitoring finance signals and determining if they are strengthened, weakened, or falsified.
Source
- Repo: https://github.com/RKiding/Awesome-finance-skills
- Path: skills/alphaear-signal-tracker
Score breakdown
Trigger
5.0
Description has What plus an explicit 'Use when monitoring finance signals...' clause; no When-Not; no trigger variants; no Chinese keywords despite China-market tools (cls/xueqiu/weibo).
Structure
6.0
51-line SKILL.md over a 7,100-line application tree (utils/schema/prompts/predictor incl. a PyTorch Kronos model) with almost no progressive disclosure; only one reference (PROMPTS.md); heavy code mass unexplained by the manifest.
Workflow
4.4
Three-step agentic workflow with role prompts is described, but the core tracking entry point is missing (SKILL.md points to non-existent track_signal method); output sanitization helper exists; test is a no-op pass; no runnable path from SKILL.md alone.
Content
5.0
Sloppy SKILL.md: duplicated '### 1. Track Signal Evolution' heading, conceptual-only example with refactor TODO comments, minimal concrete invocation guidance.
Engineering
5.0
Frontmatter valid; declared deps (agno, sqlite3) cover a fraction of actual imports (loguru, requests, jieba, torch, transformers, pandas, akshare, yfinance); referenced track_signal/FinAgent symbols do not exist in the code.
Security
8.0
Manifest: env API keys (JINA/DASHSCOPE/OPENROUTER/ZAI/DEEPSEEK) used as auth to their own providers; network to s/r.jina.ai, newsnow.busiyi.world, gamma-api.polymarket.com, dashscope.aliyuncs.com, api.z.ai, openrouter, DuckDuckGo/Baidu search, EastMoney/Yahoo, plus HuggingFace model downloads (NeoQuasar/Kronos) and local writes (sqlite db, model exports). No hardcoded secrets, no injection, no persistence mechanisms. However the description declares essentially none of this egress surface — substantive promise-capability gap (LLM routing + model downloads undeclared), major d6 deduction. No local-data exfiltration found; not veto-worthy.
A 7,100-line FinAgent application bundled as a skill behind a 51-line SKILL.md. The declared workflow (Research->Analyze->Track via references/PROMPTS.md) is sound, but the referenced implementation 'fin_agent.py track_signal' does not exist (only FinUtils), the test imports a non-existent FinAgent class and is a no-op, the example usage is an admitted 'conceptual' placeholder, and a heading is duplicated. Massive undeclared capability surface: LLM provider routing (DashScope/z.ai/OpenRouter/DeepSeek), Jina search, NewsNow, Polymarket, HF model downloads (Kronos).