Your Sunday Market Brief in One Prompt
Most people who follow markets have some version of a Sunday routine: glance at how the week closed, check whether anything big is scheduled, see which of their stocks report. It usually means five tabs and a calendar site that gives every release the same weight.
With Stocklake connected, your AI agent (Claude, ChatGPT or another) can do that round in one answer. It reads the market pulse, the current regime and outlook, a score for each of the 11 sectors, the week's macro releases and the earnings calendar, then puts them in one short brief. Below are the three prompts I use, run on Sunday, September 27, for the week of September 28 to October 2, 2026.
First: what a Sunday read actually reflects
Stocklake's market assessment and sector intelligence refresh about four times a day on weekdays, during US market hours, roughly two hours apart. Nothing runs overnight or on the weekend. So a brief you ask for on Sunday reflects Friday's last run, which in this case landed at 3:12 pm ET on Friday, September 25, shortly before the close. The pulse snapshot (VIX, fear & greed, breadth) is from the same moment.
That's fine for a weekly overview, but a good brief should say so. Each response carries an updated_at timestamp.
1. The brief itself
get_market_pulse, get_market_assessment, get_sector_intelligence, get_economic_calendar, get_earnings_calendar
The pulse
VIX read 14.97 in Friday's snapshot, down about 4.5% on the day. The fear & greed index read 37, labeled “fear.” Breadth is where the two disagree: of the 3,610 stocks with an RSI reading, 403 were oversold (11.2%) and 71 overbought (2.0%), so oversold names outnumbered overbought ones by about 5.7 to 1 (403 ÷ 71). SPY was at $771.35 with an RSI of 55.7, QQQ at $744.50 with 64.9, and small caps (IWM) at $281.97 with an RSI of 34.1.
Regime and outlook
The regime label was CAUTIOUS, with a macro_score of 39 on a 0-100 scale (0 is the most defensive reading, 100 the most risk-on). Its trend field read 0 points over 7 days and −3 over 30 days, marked “stable.” The rationale cited a calm VIX next to stress in bonds: a MOVE index of 104.6, a SKEW of 146 and the 10-year Treasury yield at 5.19%. The two things it listed to watch were an inversion of the VIX term structure (it was still in contango, 15.67 spot versus 18.43 for three months as of Sept 24) and the 10-year closing above 5.3%. The outlook label was NEGATIVE, conviction 7 of 10.
One detail worth knowing: Friday's earlier run, two hours before, had the macro_score at 33. Scores move between runs within a day, which is why the 7- and 30-day trend field is the better thing to read on a Sunday.
Sectors
Technology was the only sector with a sector_score above 50. Three sectors tied at the bottom with 18. Across the 11, the median score was 30.
| Sector | sector_score | 1 month, equal-weighted | 1 month, cap-weighted | Above 200-day average |
|---|---|---|---|---|
| Technology | 58 | +0.1% | +4.1% | 65.7% |
| Energy | 37 | −2.3% | −0.4% | 67.0% |
| Industrials | 37 | −4.2% | −1.9% | 41.0% |
| Communication Services | 26 | −3.0% | +3.5% | 41.0% |
| Consumer Cyclical | 18 | −7.3% | −5.1% | 30.3% |
| Real Estate | 18 | −7.0% | −6.6% | 24.4% |
| Utilities | 18 | −5.4% | −5.7% | 21.5% |
The two return columns tell the story of the week. Technology's cap-weighted return (+4.1%) was about 4 points ahead of its equal-weighted one (+0.1%), so a handful of large names did the lifting. Communication Services shows the same shape even more clearly: +3.5% cap-weighted against −3.0% equal-weighted. The broad market read the same way: over the month SPY was flat, QQQ up 3.3% and IWM down 5.8%.
One thing a careful answer should report rather than paper over: the sector_score_trend field, which normally gives the 7- and 30-day change for a sector, came back empty for every sector I checked, including Consumer Cyclical. The agent said so instead of inventing a direction. The only history available was Friday's own runs: Technology went 57, 58, 58, 58 through the day; Consumer Cyclical 16, 17, 17, 18.
- It says when the data is from (Friday's last run) and doesn't present it as live.
- It reports the regime and outlook as labels with their numbers, and doesn't turn them into instructions.
- It points out where the readings disagree: a calm VIX, a “fear” sentiment reading and oversold names outnumbering overbought ones more than five to one.
- It separates cap-weighted from equal-weighted performance, so a strong index doesn't hide a weak average stock.
- When a field is empty, like the sector trend here, it says so.
2. The one release that matters this week
The full calendar for a week has dozens of releases across dozens of countries. I asked the agent to pick one and give me the last reading.
get_economic_calendar, get_market_assessment
The agent picked the US jobs report: Non-Farm Payrolls on Friday, October 2 at 8:30 am ET. The prior reading was 162 (thousand jobs), with private payrolls at 127 and the unemployment rate at 4.1%. Its reasoning combined two things from the data: the calendar flags payrolls as a key event (a flag for event types that tend to move markets on their own, not a judgment on the number), and the regime rationale was built around rates, with the 10-year at 5.19% and 5.3% named as the level to watch.
It also listed the rest of the US week with prior readings: the final GDP estimate on Wednesday, Sept 30 at 8:30 am ET (prior 1.5), the ADP employment report the same morning (prior 38), and the ISM manufacturing index on Thursday, Oct 1 at 10:00 am ET (prior 54.6). Outside the US: UK GDP and a preliminary German CPI (prior 2.9% year over year) on Sept 30, and China's official manufacturing PMI (prior 49.8).
- It gives the date and time in your time zone, not just “Friday.”
- It shows the previous reading, and says plainly that the calendar carries no consensus forecast, so it can't tell you in advance what a “beat” would be.
- It explains why it picked this release, tied to something in the data, and makes clear the pick is its own reasoning.
3. Which of my holdings report?
The last part of a Sunday check is personal. I gave the agent a sample list of seven holdings: Apple, Microsoft, JPMorgan, Micron, Nike, Accenture and Carnival.
get_earnings_calendar, get_stocks, get_sector_intelligence
| Holding | Reports | Sector | sector_score |
|---|---|---|---|
| CCL | Tue Sept 29, before the open | Consumer Cyclical | 18 |
| MU | Wed Sept 30, after the close | Technology | 58 |
| ACN | Thu Oct 1, before the open | Technology | 58 |
| NKE | Thu Oct 1, after the close | Consumer Cyclical | 18 |
| JPM | Tue Oct 13 (next week) | Financial Services | 30 |
So four of the seven report this week, two in the highest-scoring sector and two in one of the lowest. Apple and Microsoft don't report this week; their next dates hadn't been published yet, and the answer said that instead of guessing. The wider calendar had 18 companies from Stocklake's universe reporting between Sept 28 and Oct 2, five of them in Consumer Cyclical and five in Technology. The largest by market value were Micron, Accenture, Nike, Jabil and Carnival.
- It gives the day and whether the report comes before the open or after the close.
- It says whether each date is confirmed or an estimate.
- For holdings without a date this week, it says “not scheduled yet” rather than filling in a guess.
- It connects each holding to its sector's reading, so you see which reports land in a strong or weak part of the market.
Free vs Pro for this routine
A useful part of the brief works on the free tier. The regime, the sector scores and the full calendar are where Pro comes in.
- Market pulse: VIX, fear & greed, breadth, SPY/QQQ/IWM
- The earnings calendar, with dates and confirmed-or-estimate
- Sector, price and next earnings date for any stock
- Screener and market movers
- Market assessment: regime, outlook, macro_score and its trend
- Sector intelligence: sector_score, breadth and movers for all 11
- The full economic calendar, with key-event flags and prior readings
- AI verdict and ai_score for each holding
Tips for a better brief
- Ask for the timestamp. “Tell me when each piece of data is from” keeps a Sunday answer honest about being Friday's data.
- Ask your agent to cite the fields it used. “Show the numbers behind each sentence” gets you macro_score 39 and breadth 11.2% oversold instead of “the market looks nervous.”
- Save the prompt. The same three prompts work every weekend. Asking “what changed since last Sunday?” in the same chat is a good fourth one.
Setting it up
Stocklake works in Claude, ChatGPT and other MCP clients. You add it once as a connector, using the server URL https://api.stocklake.dev/mcp, and it's then available in every chat. You can start on the free tier without a card and add Pro later. Step-by-step setup for each client is in the docs.
Frequently asked questions
Stocklake's market assessment and sector intelligence refresh about four times a day on weekdays, during US market hours. Nothing runs over the weekend, so a brief you ask for on Sunday reflects Friday's last run. The response includes an updated_at timestamp, and a good answer quotes it.
Yes, a partial one. The free tier includes the market pulse (VIX, fear & greed, breadth, SPY, QQQ, IWM) and the earnings calendar, plus the sector and earnings date for any stock. The market assessment with macro_score, sector intelligence with sector_score, and the full economic calendar are Pro.
No. CAUTIOUS or NEGATIVE are labels the data returns, summarizing volatility, breadth, rates and sector strength. They describe conditions. Stocklake doesn't generate trade calls, and nothing in this post is a recommendation to buy, sell or short any security.
Yes, as long as your AI app supports MCP connectors in its mobile app. Once Stocklake is added as a connector to your account, it's available wherever you use that account: on the web, on desktop and on your phone.