Day Trader Skill: Install, Paper Trading Defaults, and How to Audit Before Live Use

Day Trader Skill is an MIT Claude skills pack for day-trading workflows. Install paths, paper-trading defaults, risk gates, and how to audit SKILL.md files and network calls—without stock tips or return promises.

📊 Browse AI app rankings

📱 Get the AI Rank app

Day Trader Skill is an open MIT pack of 11 Claude skills for day-trading workflows. The repo is github.com/Jakeschincariol/day-trader-skill (canonical casing Jakeschincariol/day-trader-skill). It reads free public market data with plain Python, defaults to paper trading, and states that no broker order goes out until you say yes. This article explains what is in the pack, how to install it, how paper and risk defaults work, and how to audit each SKILL.md and network call before you connect a live broker. Checked November 4, 2026 (America/Los_Angeles).

Hard limits for this page: we do not recommend stocks, promise returns, or claim we ran live P&L. We did not paper-trade or live-trade with this pack for this article. Day trading can lose money. Treat the README's own risk language as the floor, not optional flavor text.

What Day Trader Skill is

The README describes eleven skills that chain through a local trader/ folder: chart scanner, news monitor, entry timer, exit planner, time-machine simulator, risk manager, budget / position sizing, broker ticket executor, price alerts, earnings analyst, and a Congress PTR ("Pelosi") tracker.

Data sources named in the README: Yahoo Finance charts, Nasdaq Trader symbol directories, SEC EDGAR, and the House Clerk. Tools are standard-library Python with nothing to pip install. License: MIT. The project says free use, no signup, and no API key for the core data tools (SEC tools ask you to identify yourself with a user-agent string).

The README positions this repo as the day-trading sequel to Jakeschincariol/trader-skill ("Trader skill v2"). There is no root SKILL.md at the repository root; each skill lives in its own trader-*/SKILL.md folder.

Stars and activity change. When checked for this article the GitHub API reported thousands of stars and a default branch of main. Re-check the live repo page.

For other Claude skill packs on AI Rank, see Replica Skill and the Capafy AI skills marketplace notes. Those are separate projects.

How to install Day Trader Skill

README install paths:

Paste into Claude: send the repo URL and the words install skill.
Claude Code plugin:
- /plugin marketplace add Jakeschincariol/day-trader-skill
- /plugin install day-trader-skill@day-trader-skill
- Plugin-namespaced commands look like /day-trader-skill:trader-scanner.
Copy folders for plain /trader-scanner names:
- git clone https://github.com/Jakeschincariol/day-trader-skill.git
- cp -r day-trader-skill/trader-* ~/.claude/skills/
- Or copy into a project-local .claude/skills/.
Chat-only fallback: paste a single skill's SKILL.md into a chat. The README says you lose the Python tools but keep the method text.

Python 3.8+ is required for the scripts. For SEC-facing tools, set "sec_user_agent": "Your Name you@example.com" in trader/config.json or the SEC_USER_AGENT environment variable.

The eleven commands (as listed in the README)

Command	Role (README)
`/trader-scanner`	Scan charts for setups (opening-range breakouts, flags, oversold bounces)
`/trader-news`	Track breaking headlines / filings on watched tickers
`/trader-entry`	Watch a setup bar by bar for an entry trigger
`/trader-exit`	Set stop and target before entry
`/trader-timemachine`	Replay simulated session outcomes from recent bars / past setup hits
`/trader-risk`	PASS / BLOCK orders against configured limits
`/trader-budget`	Size positions to budget and max loss per trade
`/trader-broker`	Show tickets and send only after your yes (paper by default)
`/trader-alerts`	Alert on price, stop, or flatten time
`/trader-analyst`	Summarize earnings / XBRL-style facts from public filings
`/trader-pelosi`	Read public congressional PTR disclosures

The README's example pipeline for one trade is: scanner → news → entry → exit → timemachine → budget → risk → broker, with alerts while a trade is open.

Paper trading default and risk controls

From trader-budget/config.example.json in the repo:

"paper": true is the default.
Example budget fields include budget, max_loss_per_trade_usd, max_daily_loss_usd, max_trades_per_day, require_stop, entry window times, flat_by, and flags that default off for options, margin, and shorting (allow_options, allow_margin, allow_short are false in the example).
pdt_guard defaults to false; the README notes FINRA's pattern-day-trader rule changes and tells you to enable the guard if your broker still enforces an older minimum.

trader-broker/SKILL.md states three gates before any market order: trader-risk PASS for that exact order, an explicit yes in chat for that order, and Claude's Allow once permission card because the broker tool must stay on Needs approval (never Always allow). The skill text says ticket.py itself never talks to a broker; it renders tickets and records approvals. Alpaca's official MCP server is documented as paper by default, with keys only in local config / env vars, never in chat.

trader-risk/SKILL.md lists checks that can BLOCK entries (format, asset type, shorts, blocked tickers, size, position %, cash buffer, stop required, max loss, reward:risk, daily loss, trades per day, entry window, optional PDT guard, duplicates, daily order cap). Exits are described as never blocked by day-loss style rules so you can flatten.

README fine print you should keep:

Not financial advice; most people who day trade lose money.
Time Machine odds are a simulation from recent bars, not a prediction.
Stops can slip; max loss shown is loss at the stop price.
Yahoo chart data is unofficial and can lag or rate-limit.
Congressional trades are public STOCK Act disclosures, not insider tips.
Tax treatment of day trades is on you; the pack is not tax advice.

How to audit SKILL.md files and network calls before use

Do this before pointing the pack at a funded account:

Read every trader-*/SKILL.md you will enable. Confirm the description matches the behavior you want. Pay special attention to trader-broker/SKILL.md (order path) and trader-risk/SKILL.md (what BLOCK actually covers).
Diff config.example.json into your own trader/config.json. Keep "paper": true until you have watched PASS/BLOCK behavior on fixtures.
Run the offline tests the README documents: python3 -m unittest discover -s tests -v. The suite uses fixtures cut from real historical data (the README names TSLA bars around October 19, 2026 among them). Passing tests does not prove live profitability.
Inventory network destinations. Scripts talk to public endpoints named in the README (Yahoo Finance, Nasdaq Trader directories, SEC, House Clerk) and, only if you configure it, a brokerage MCP such as Alpaca. Open the Python files for the skills you enable and note every URL host before first run. Block or avoid any host you do not accept.
Keep broker connectors on manual approval. README: set order tools to Needs approval in Claude connector permissions.
Do not paste live API secrets into chat. Keys belong in env vars or local config files excluded from git.
Assume marketing pages outside GitHub are unofficial. Third-party roundups (for example skills directory sites) can lag or misstate defaults. Prefer the repo README and the SKILL.md files you cloned.

If a skill folder's SKILL.md and its .py tool disagree, trust neither blindly: open both, or disable that skill until you understand the call path.

FAQ

Is Day Trader Skill free?

The GitHub project is MIT licensed. The README says no signup and no API key for the core public-data tools. Broker accounts and any paid Claude plan are separate.

Does it place live trades by default?

No. Paper trading is the documented default. Only /trader-broker can lead to an order, and only after risk PASS plus your explicit yes plus Claude's Allow once card, per the broker skill text.

Did AI Rank test this for profit?

No. This article is a documentation read of the public repo. We did not measure win rate, expectancy, or account P&L.

Is this financial advice?

No. It is software documentation. Decisions about money stay with you and any licensed advisor you choose.

Where is the root SKILL.md?

There is no repository-root SKILL.md. Skills are packaged per folder under trader-*.

Sources

GitHub: Jakeschincariol/day-trader-skill (README: install, eleven skills, paper default, fine print; MIT LICENSE; checked Nov 4, 2026)
Repo files: trader-broker/SKILL.md, trader-risk/SKILL.md, trader-budget/config.example.json, and sibling trader-*/SKILL.md paths listed in the README Files section
Related AI Rank explainers (separate projects): Replica Skill, Capafy AI

Not affiliated with the day-trader-skill maintainers, Anthropic, Alpaca, or Yahoo. Repo contents and broker APIs change; re-read the files you install before any live use.

Keep exploring: see other Claude skill write-ups like Replica Skill and Capafy AI, or browse the open-source ranking on AI Rank.

Or get the AI Rank app.