1. Connect the Bot and keep writes behind approval
In Grok Bot, open Marketplace in the sidebar. If a NexusTrade connector is available, select Add, complete its browser authentication and attach it with @ in this task. For your existing or custom MCP connection, give the Bot the current NexusTrade endpoint below and ask it to complete OAuth discovery. This is a connection request; verify the returned tools and account before continuing.
For sign-in or two-factor authentication, open Agent Computer, take control, finish the sensitive step and return control. Browser sessions and command-line credentials are shared by your account's Bots. Keep the experiment notes in a named folder under /workspace.
Open Settings → General → Auto-review and add Ask first rules for paid research, backtests, portfolio changes and deployment. Review the exact inputs on the approval card and use Allow once for a single accepted operation. The published NexusTrade case study documents Grok Bot MCP and Python SDK use. The SDK guide provides a terminal route when the current connector cannot finish authentication; its API key is separate from MCP OAuth.
Connect this Grok Bot task to NexusTrade MCP at https://nexustrade.io/api/mcp using OAuth discovery. Complete the sign-in handoff with me, list discovered tools, then run only the account read check. Do not launch research or backtests, activate a portfolio or place orders.2. Check the account with an owned portfolio read
Ask the agent to call fetch_portfolios with the arguments below. Compare the returned portfolio names, IDs and types with your NexusTrade account. fetch_portfolios returns portfolioId; pass that exact value as portfolio_id to get_portfolio. Inspect the structured portfolio.portfolioId, portfolio.type, portfolio.isActive, portfolio.strategies, portfolio.initialValue and portfolio.deployment. For a new account, an authenticated empty list is a valid result. Public stock commentary does not establish account access.
This request excludes live books and positions, includes chat drafts and inactive paper books, and returns up to 50 records per page. Follow subsequent pages before concluding a named book is absent. If OAuth expired, reconnect and repeat this read. For a missing tool, inspect the included app, environment and action permissions. An account mismatch must be fixed before a write.
{
"include_live": false,
"include_paper": true,
"include_chat_portfolios": true,
"include_inactive": true,
"include_positions": false,
"limit": 50,
"page": 1
}3. Build the NVDA brief from the company release
NVIDIA published fiscal Q2 2027 results on August 26, 2026 for the quarter ended July 26. Its GAAP table reports $96,221 million of company revenue versus $46,743 million a year earlier; the release rounds the current figure to $96.2 billion and reports 106% growth. Data Center revenue of $89.0 billion is a segment figure, not the company total.
Use the Bot's saved notes to separate reported revenue from management's next-quarter outlook. Revenue growth alone cannot answer whether receivables, customer concentration or supply commitments make that growth sustainable. Ask for the corresponding 10-Q and retain its publication date before interpreting those risks.
| Issuer and ticker | NVIDIA · NVDA |
| Reporting period | Fiscal Q2 2027 ended July 26, 2026 |
| Published | August 26, 2026 |
| GAAP revenue | $96,221 million; prior year $46,743 million |
| Reported growth | 106% year over year |
| Segment distinction | $89.0 billion Data Center revenue is part of company revenue |
No matching rows. Clear the filter to see all records.
4. Read the filing list before starting paid research
Use sec_edgar to list NVDA filings with the read-only request below. Open the relevant 10-Q and earnings 8-K; confirm issuer, accession, reporting period and filed date. The example above is a dated source check, not a claim that it will remain the latest quarter.
Missing filing values stay unresolved. action=ingest_filings persists source text and is a separate write, so listing filings does not grant permission to ingest them.
{
"ticker": "NVDA",
"action": "list_filings",
"forms": [
"10-Q",
"8-K"
],
"since": "2026-07-01",
"limit": 10
}Read the reported quarterly fundamentals
Use this separate sec_edgar request to read available quarterly XBRL metrics. Match each observation to its fiscal period, unit and filing date rather than assuming the first row is the quarter in this example. Open the issuer statement to resolve any discrepancy. A missing operating cash-flow value remains missing.
{
"ticker": "NVDA",
"action": "fundamentals",
"periods": "quarterly"
}5. Submit one bounded question, then poll its ID
After reviewing the question and its usage cost, explicitly approve one research_topic call with message. The structured result returns researchJobId and status. Copy that returned ID into get_research_status as research_job_id. PROCESSING means poll the same job; COMPLETED should include the report and citations; FAILED includes an error to inspect.
If the response is lost, inspect the task transcript for the accepted ID before starting another request. A second research_topic call can create another billed job. A missing or foreign ID is an ownership or identity problem, not an instruction to invent a result. Read each primary citation before accepting the brief.
After I approve one research_topic request, use message: "Research NVDA using the primary earnings release and corresponding 10-Q. Compare reported revenue and operating cash flow with the matching prior-year period. Preserve GAAP basis, units, quarter-end and publication dates. Explain disclosed one-off items, customer concentration and missing evidence. Return facts, interpretation and disconfirming evidence with source URLs. Do not draft orders or alter portfolios."
Save the returned researchJobId. Call get_research_status with research_job_id set to that exact value until terminal status.From a question to evidence
research_topic is asynchronous. Poll get_research_status with the same job ID instead of launching another billed research run for the same question.
| Read filings | sec_edgar | Issuer, accession, reporting period and publication date |
| Read relevant news | search_stock_news | Dated publisher links; distinguish news from filings |
| Submit bounded research | research_topic | researchJobId and source-backed question |
| Poll the accepted job | get_research_status | Same research_job_id, completed report and citations |
No matching rows. Clear the filter to see all records.
6. Save the evidence and the rejected thesis in the Bot workspace
Keep the NVDA source ledger, completed report and accepted job ID in /workspace/nvda-research. Ask the Bot to reopen those files when resuming. Its workspace persists, but a saved market claim still needs a dated source check.
Austin's published Grok Bot article records MCP and SDK strategy work, followed by a human-directed allocation redesign and manual approval of live orders. Use that as evidence that the workflow was used. Its historical tests and later account statements are dated case-study observations, not measured results for the new NVDA assignment.
Save the NVDA brief, source URLs, publication dates, researchJobId and open questions in /workspace/nvda-research. Preserve rejected interpretations and their evidence. On resumption, read the saved job record before making another paid request. Ask before creating or changing a portfolio.Choose a separate strategy experiment
A filing brief ends with a documented thesis and what would change it. Turning that thesis into a rule requires a separate entry, exit, sizing and timing decision. The price-rule guide below tests mechanics; it does not retrospectively treat a 2026 earnings release as information known in 2021.
