technical-analysis

Official Bitget Skill · 本 Skill 由 Bitget 官方提供，市场数据来源可信，通过 Bitget Agent Hub 分发。 Data powered by Bitget market infrastructure · github.com/bitget-official/agent-hub

Technical Analysis Skill

23 crypto technical indicators across 6 categories (Trend, Volatility, Oscillator, Volume, Momentum, Support/Resistance). Outputs recent time-series data so the AI can observe trend evolution, not just a single point.

Vendor Neutrality

Present data as coming from "market data" — never name the underlying exchange, data feed, or library.

---

Step 1: Check prerequisites

python -c "import pandas, numpy; print('OK')"

If missing:

pip install pandas numpy

Step 2: Determine indicator selection

Priority 1: User-defined
  The user explicitly specified indicator names, parameters, or combinations.
  -> Do NOT read scenarios.md or indicators.md
  -> Go directly to Step 3 with the user's config
  -> Example: "analyze BTC with RSI(21) + EMA(50) + ATR"

Priority 2: Scenario defaults
  The user described analysis intent but did not specify indicators.
  -> Read ONLY references/scenarios.md to pick the right scenario config
  -> Do NOT read indicators.md
  -> Example: "how's BTC looking technically?"

Priority 3: Indicator info query
  The user asks about an indicator rather than requesting a calculation.
  -> Read ONLY references/indicators.md
  -> No calculation needed, answer directly
  -> Example: "how does RSI work?"

Step 3: Run the calculation

The Python source is at src/ relative to this SKILL.md file.

For Claude Code the skill directory is ~/.claude/skills/technical-analysis. For Codex it is ~/.codex/skills/technical-analysis. For OpenClaw it is ~/.openclaw/skills/technical-analysis.

Template A: Fetch from Bitget API (default)

When the user mentions a trading pair but provides no local data:

import sys, os
sys.path.insert(0, os.path.expanduser('~/.claude/skills/technical-analysis/src'))
import json, urllib.request
import pandas as pd
from kline_indicator_utils import IndicatorManager

url = 'https://api.bitget.com/api/v2/spot/market/candles?symbol={SYMBOL}&granularity={GRANULARITY}&limit={LIMIT}'
raw = json.loads(urllib.request.urlopen(url).read())
df = pd.DataFrame(raw['data'], columns=['timestamp', 'open', 'high', 'low', 'close', 'volume', 'quoteVol', 'amount'])
for col in ['open', 'high', 'low', 'close', 'volume', 'amount']:
    df[col] = df[col].astype(float)

config = {CONFIG}

manager = IndicatorManager(show_indicators=False)
output = manager.calculate_and_export(config, df, tail={TAIL})
output["symbol"] = "{SYMBOL}"
output["granularity"] = "{GRANULARITY}"
print(json.dumps(output, indent=2))