Repository navigation
Skill Proposal: nse-bse-hybrid-trader (Indian NSE/BSE Trading Skill with Stdlib Risk & Monte Carlo Engine)[Idea] #2304
aryanpardeshi11
started this conversation in
Ideas
Replies: 1 comment 1 reply
|
Thanks for putting this together, @aryanpardeshi11. Please open a draft PR with the complete bundle, including the scripts, and link back to this discussion. Include the offline and mocked test results you mentioned so I can review the implementation and the paper and live trading paths. |
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
What problem are you trying to solve?
There is no Indian NSE/BSE skill in the catalog that pairs multi-timeframe technical analysis with deterministic 2026 statutory tax math, INR/lot sizing, Monte Carlo validation, and paper-first broker safeguards so agents do not hallucinate financial arithmetic.
Proposed idea
Add the MIT-licensed
nse-bse-hybrid-traderskill bundle (skills/nse-bse-hybrid-trader/) for Indian equities, indices (NIFTY,BANKNIFTY,FINNIFTY,SENSEX), and F&O.Because
CONTRIBUTING.mdrequires maintainer sponsorship before submitting restrictedscripts/files via fork PRs, I am describing the complete bundle here first:0mandatorypipruntime packages):SKILL.md+references/technical_frameworks.md&references/indian_market_rules.md: 3-screen MTF workflow,0–6confluence rubric, RSI divergences, Fibonacci Golden Pocket, SEBI ASM/GSM & F&O ban rules, and## When to Use/## Limitations/ safety guardrails (descriptionkept< 200chars perquality-bar.md,category: finance,risk: critical).config/trading_rules.json&config/nse_holidays.json: April 2026 Indian statutory tax rates (STT, NSE fees, SEBIRs. 10/Cr, stamp duty, DP charge, 18% GST), 9 broker fee tiers, risk caps, and 2026 NSE/BSE holiday calendar.scripts/nse_common.py&scripts/nse_data_loader.py:DecimalRs. 0.05tick math, IST windows, pure-Python EMA/SMA/RSI/MACD/Bollinger/ADX/ATR/VWAP, 15m ORB, Fib swings, divergences, and Yahoo v8 /--offline-*loaders.scripts/nse_timemachine.py: Deterministicseed=1910,000-path time-of-day bootstrapped Monte Carlo (15:15 ISTsquare-off cutoff, stop-first checks, 1-tick slippage) + historical 15m ORB replay.scripts/indian_risk_gate.py: Tax-inclusive INR sizing,2%risk ceiling,1.5Net R:R floor,3%daily drawdown lock reading local SQLite, and SHA-256 verdict signing.scripts/broker_bridge.py+scripts/smartapi_bridge.py: SQLite ScripMaster cache, default--paperledger (state/paper_trades.db, zero external broker calls/keys), and gated--live --confirm "CONFIRM LIVE"REST clients (urllib.request) for 9 Indian brokers (Zerodha, Angel One, Upstox, Dhan, Fyers, Groww, Kotak Neo, ICICI Breeze, Shoonya). Note:_strip_attached_fields()strips bracket fields on non-bracket live orders.LICENSE& Exclusions: Retainsskills/nse-bse-hybrid-trader/LICENSE(MIT); excludes runtimestate/,.env,*.db, and.claude-plugin/.12/12offline unit & mocked 9-broker HTTP tests pass viapython -m pytest -v.Real use cases
1.714R) looks valid,nse_timemachine.py(seed=19, 59 full 5m NSE sessions,15:15 ISTcutoff) measures17.94%target-first odds (-0.2893Rnet EV), andindian_risk_gate.pycomputesRs. 57.402026 taxes/fees (7.59%drag), capping size at63shares, lowering Net R:R to1.404 < 1.50, and issuing a SHA-256 signedBLOCK(NET_RR_TOO_LOW) thatbroker_bridge.pyrefuses to stage.--paperdefault, zero demat/API keys): New traders stage entries into localstate/paper_trades.db, check daily drawdown lock, and settle tax-adjusted Net P&L without connecting real broker credentials.NIFTY,BANKNIFTY, orRELIANCE.NSacross Weekly/Daily/5m–15m alignment, VWAP, 15m ORB, Fibonacci pockets, and SEBI surveillance constraints (--fno-ban,--asm-gsm).Related skills, tools, or prior art
quant-analyst(general financial models),traderspy-trading-signals(crypto futures signals),trading-ledger(Notion trading journal) — none cover NSE/BSE rules, 2026 Indian taxes, or Indian brokers.Are you willing to help build this?
Yes, I can open a PR
Before posting
All reactions