agentleFS
ShubhamkumartrichyShubhamkumartrichy/claude-trading-skills-india/skills/portfolio-review/SKILL.md

portfolio-review

Skill64 starsChanged 17 days agoMITView on GitHub

Health check for an Indian delivery (CNC) stock portfolio from a holdings CSV or a Zerodha Console, Kite, Groww or broker-MCP export: value and unrealised P&L, position and sector weights, top-5/top-10 concentration, business-group exposure (Adani, Tata...), large/mid/small mix against a target, pledged, SME and surveillance flags, a per-lot STCG→LTCG countdown, volatility, beta vs NIFTY, drawdown, correlated clusters, portfolio heat to stops, and a tax-aware rebalance plan with charges.

15 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Look at the whole book before the next trade.

    line 18
  2. Very specific:

    Every command takes --as-of YYYY-MM-DD and --json.

    line 32
  3. Very specific:

    Prefer --no-sell (new SIP or bonus money into underweights) when selling would realise STCG.

    line 102

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Portfolio Review (India)

Look at the whole book before the next trade. This skill reports facts and flags (INFO / WARN / REVIEW). It never says "sell": every flag is a prompt for the user to re-check a thesis.

Commands

python3 scripts/portfolio.py review holdings.csv --cash 200000                    # weights, flags, tax clock
python3 scripts/portfolio.py review console.xlsx --sheet Equity --max-stock 8 --mcap-target 50/30/20
python3 scripts/portfolio.py risk holdings.csv --prices-dir prices/ --benchmark NIFTY    # or --yf
python3 scripts/portfolio.py rebalance holdings.csv --cash 50000                   # equal weight
python3 scripts/portfolio.py rebalance holdings.csv --targets targets.csv --broker groww --realized-ltcg 40000
python3 scripts/portfolio.py rebalance holdings.csv --cash 75000 --no-sell         # deploy new money only

Every command takes --as-of YYYY-MM-DD and --json. --prices-dir holds daily candles named SYMBOL.csv or SYMBOL.json (any format ../technical-analysis/scripts/ta.py reads) plus NIFTY.csv. Targets file: symbol,weight[,price] or sector,weight, in percent. Weight 0 means exit; symbols or sectors left out stay unchanged. A price column lets you add a stock you do not hold yet.

Input columns

Required: symbol, qty, buy_price (average cost), ltp. Strongly recommended: buy_date, with one row per lot, as in ../indian-tax-for-traders/scripts/tax.py harvest files. Optional: sector, mcap (large/mid/small, or a number in ₹ crore), group, stop, pledged_qty, series (EQ/BE/SM...), sme, illiquid, surveillance (ASM/GSM text), promoter_pledge (%), isin, exchange, fmv_2018.

Broker headers are mapped automatically, and preamble rows above the header are skipped:

SourceHeaders recognisedStatus
Kite Connect / Kite MCP get_holdingstradingsymbol, quantity + t1_quantity, average_price, last_price, collateral_quantity, isin, exchangeChecked against the Kite Connect v3 docs
Zerodha Console holdings (XLSX)Symbol, ISIN, Sector, Quantity Available, Quantity Long Term, Quantity Pledged (Margin/Loan), Average Price, Previous Closing PriceAssumed. Compare with your file
Kite web holdings CSVInstrument, Qty., Avg. cost, LTP, Cur. valAssumed
Groww holdings statement (Reports)Stock Name, ISIN, Quantity, Average buy price, Buy value, Closing price/valueAssumed. Gives company names, so add a symbol column before running risk --yf

Broker MCPs (broker-integration): call the holdings tool and write a CSV with one row per holding using the fields above. Keep the raw field names; the mapper understands them. Broker exports carry no buy dates. Console's "Quantity Long Term" is still used, but for a lot-level countdown, rebuild open lots with tax.py tradebook and add buy_date. Keep holdings files local. Do not commit them.

Defaults and why

SettingDefaultReason
--max-stock10% of capitalOne bad result or fraud should not sink the book. REVIEW above 2×
--max-sector25%Sectors move together on policy, commodity and rate news
--max-group20%Group companies share promoters, lenders and governance risk
--mcap-target60/25/15A core-heavy mix for a retail investor. Change it to match your plan
--mcap-cutoffs₹1,00,000 cr / ₹33,000 crRough stand-ins for AMFI's rank-based list (top 100 / 101–250). Update every half-year
--loser-pct20% below costA decline that big needs a thesis check
--max-heat6% of capitalThe total loss if every stop is hit together
--band / --min-trade1 pp / ₹1,000Avoids churn where DP charges (~₹15 per scrip per sell day) and STT eat the benefit

Holdings without a sector use a built-in map of about 55 NIFTY 50 names, marked *, which are also treated as large caps. Anything else shows as Unknown, so add the column.

Reading the results in an Indian context

  • Your pledge vs promoter pledge. pledged_qty is shares you pledged for F&O margin. A fall shrinks collateral and can trigger forced square-offs. promoter_pledge above 20% is a governance red flag; see fundamental-screening.
  • ASM/GSM/ESM and BE/BZ series. These bring higher margins, trade-for-trade settlement and tight price bands. Exits get slow just when you need them.
  • SME stocks (NSE Emerge, BSE SME; series SM/ST) trade in large lots with thin books. Size them as illiquid, whatever their weight.
  • Group concentration. Three Adani or Tata companies at 8% each fit under the single-stock limit but form one 24% bet on a single promoter group. Add a group column; a small built-in map covers common NIFTY names.
  • Correlated clusters from risk, such as private banks, PSU banks or IT exporters, are where diversification is weaker than the stock count suggests.
  • Beta above about 1.2 means a 10% NIFTY fall is likely to cost more than 10%. With cash, the "incl. cash" beta is the one that matters.

Tax-aware rebalancing

rebalance sells oldest lots first (FIFO, as for demat shares) and takes the holding-period rules, grandfathering and rates from indian-tax-for-traders (tax.py). Charges come from trading-costs (../trading-costs/scripts/charges.py). For each sale it flags:

  • WAIT? A profitable short-term lot turns long-term within --wait-days (default 60). Selling now is taxed at 20%; waiting brings it to 12.5%, or nil inside the ₹1.25 lakh exemption. Weigh the tax saved against the price risk.
  • LTCG inside the exemption. Pass --realized-ltcg so the remaining exemption is right.
  • Booked losses. Short-term losses offset STCG or LTCG; long-term losses offset only LTCG.

Prefer --no-sell (new SIP or bonus money into underweights) when selling would realise STCG. For harvesting ideas, run tax.py harvest on the same file; review already lists them.

How it fits with other skills

  • market-regime sets how much equity exposure makes sense. portfolio-review checks what you hold. In a risk-off regime, look at heat and beta first.
  • sector-rotation says which sectors lead. Feed its view into a sector,weight targets file.
  • position-sizing and stop-loss-strategies supply the stop column that makes heat meaningful.
  • trading-discipline runs the pre-trade gate. Treat heat above the limit, or a WARN on the stock or sector you are adding to, as a reason to stop at that gate.

Output

Portfolio review as of 2026-11-01 (6 stocks, 7 lots)
  Holdings ₹8,00,000 | Cash ₹2,00,000 (20.0%) | Total ₹10,00,000 (₹10.00 L)
  Invested ₹7,10,000 | Unrealised P&L ₹90,000 (+12.68%) | Top-5 76.0%
  Sectors (limit 25%): Information Technology 55.0%, Energy 10.5%, ...
  Tax clock: 1 lot turns long-term within 30 days, 1 in 31–60 days
  REVIEW AAA is 30.0% of capital (single-stock limit 10%).

Lead with the two or three flags that matter most, then the numbers. Say which inputs were missing (sector, buy dates, prices). Never turn a flag into an order. Orders follow the broker-integration confirmation protocol.

Educational analysis, not investment or tax advice. Data: [holdings file + as-of date].

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.