Menu
Turn a news headline or event into a well-formed prediction market
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Create, cancel, or manage multiple orders in a single batch
Place and manage prediction market orders on Context Markets
Build an embeddable single-market prediction widget with buy/sell buttons
Build a portfolio dashboard showing positions, P&L, balances, and claimable winnings
Build prediction market frontends with the Context React SDK
Scaffold a full prediction market trading UI with market list, orders, and portfolio
Based on SOC occupation classification
| name | context-create-news-to-market |
| description | Turn a news headline or event into a well-formed prediction market |
Transform a news article, headline, or event into a properly structured prediction market with clear resolution criteria.
The user has a news event or topic and wants to create a prediction market around it.
Evaluate the input — is the event resolvable, binary, timely, and interesting?
Identify the claim type — event-by-deadline, threshold, period-gated, durational, or none/never. This determines how the oracle can resolve.
Draft the question — start with "Will...", include a specific outcome and timeframe. Keep under 150 characters. Create a shorter shortQuestion variant.
Write resolution criteria — the most important part:
This market resolves YES if [specific condition].
This market resolves NO if [the market end time passes without the condition being met].
Evidence sources: [@handle1, @handle2] on X/Twitter.
[OR] Evidence sources: Official announcements from [specific outlets].
Clarifications:
- [What counts/doesn't count]
- [Edge cases]
Choose evidence mode:
social_only if X/Twitter accounts will cover the eventweb_enabled if official data sources (SEC filings, sports scores, government sites) are neededList sources — for social_only: specific X handles. For web_enabled: types of authoritative sources.
Set end time — give buffer after the expected event for evidence to appear. If the event happens at 3pm, set end time to midnight or later.
Submit:
context_agent_submit_market({ formattedQuestion, shortQuestion, marketType: "OBJECTIVE", evidenceMode, resolutionCriteria, endTime: "2026-06-15 23:59:59", timezone: "America/New_York", sources: ["@handle"], explanation: "Brief reason" })ctx.questions.agentSubmitAndWait({ market: { formattedQuestion, shortQuestion, marketType, evidenceMode, resolutionCriteria, endTime, timezone, sources } })ctx.markets.create(submission.questions[0].id)context questions agent-submit-and-wait --formatted-question "..." --short-question "..." --market-type OBJECTIVE --evidence-mode web_enabled --resolution-criteria "..." --end-time "2026-06-15 23:59:59" --timezone "America/New_York" --sources "@Apple,@tim_cook"Verify — call context_get_market with the returned market ID to confirm it was created.
social_only, or describe authoritative sources for web_enabled.social_only for an earnings report that will be on the SEC website but not tweeted by anyone specific. Use web_enabled for official data.context_get_market returns the created market with status "active".