Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a non-advisory prediction-market trade planning worksheet for Itô or venue workflows. Use to inspect venues, underliers, constraints, order prerequisites, and manual execution steps without placing trades or recommending positions.
.claude/skills/ito-trade-planner/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 123% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -42% | 0% |
Use this skill when a user wants a structured worksheet for a prediction-market idea, basket adjustment, venue comparison, or manual execution plan.
The skill is intentionally non-executing. It produces indicative, non-executable checklists and parameter tables the user can review manually.
credentials.
research to execution-capable tooling. This approval does not authorize this skill to execute anything.
or opening an execution-capable tool or venue.
The canonical developer surface is https://itomarkets.com/api/v1. Use only authenticated GET endpoints requiring baskets:read or markets:read, either with HTTPS and Authorization: Bearer $ITO_API_KEY or the official ito-markets Python SDK. Trading is not part of this API.
On first use, check for an already configured key with exactly baskets:read and markets:read without printing it. Least-privilege public keys use the bkt_* form and are operator-issued; the dashboard's Settings -> Keys & credentials flow issues a broader ito_* automation key. Do not create or rotate that broader key merely to unblock this skill. If a scoped key is unavailable, report the read-only API route as blocked and continue with clearly labeled public or user- supplied inputs. Key issuance creates persistent access and needs confirmation in the controlling harness. After the user or operator stores the one-time value securely, return control to the originating agent and run one minimal GET /baskets auth probe. This API does not use device authorization or device login; do not invent a verification-code handoff.
The ecc ito bridge is a separate compute-procurement surface. Do not use ecc ito login, ecc ito find, or its MCP tools for prediction-market data or trade planning. Never print, log, persist, or place ITO_API_KEY in arguments, reports, screenshots, tracked files, or chat. Retrieve only the minimum field at runtime and keep it in process memory.
Mark API observations indicative. Use GET /baskets, GET /baskets/{basket_id}, GET /baskets/{basket_id}/price, GET /baskets/{basket_id}/underlyers, GET /markets/search, and GET /markets/{market_id} as needed. Do not use write or backtest submission endpoints for a trade-planning worksheet.
freshness constraints.
record the endpoint URL and retrieved_at timestamp. Never infer a live price from stale, missing, or inaccessible data; use unknown.
eligibility, venue, market identifier, side (if the user supplied one), limit, time-in-force, maximum spend, fees, liquidity/slippage boundary, resolution rule, and decision deadline. Missing constraints remain unknown.
prediction-market-risk-review before discussing automation, keys,venue auth, capital constraints, or a manual action link.
request separate explicit confirmation in the future execution-capable workflow. Do not treat confirmation given during planning as an order.
401, set plan_status: blocked and ask the user to inspect or replace thekey in Settings. On 403, report the missing read scope; never request a write scope for this skill. Redact any credential-like text.
429, honor Retry-After once within the user's time budget. Do not loop orexceed the documented read budget of 120 requests per minute.
unknown.Retry at most once for a read; never turn a read failure into a write.
credentials, and direct the user to Settings. Never weaken scopes or reuse cached secrets.
or fixture data as live behavior.
Use:
Avoid:
Return this shape in Markdown or YAML. Preserve unknown rather than guessing.
yamlplan_status: ready_for_manual_review | blocked mode: indicative_non_executable hypothesis: "neutral restatement" markets: - market: "identifier or unknown" venue: "venue or unknown" observable_status: "value or unknown" source_url: "source URL or unknown" retrieved_at: "ISO-8601 timestamp or unknown" resolution_rule: "summary or unknown" liquidity_caveat: "text or unknown" constraints: jurisdiction_eligibility: "confirmed | unconfirmed | unknown" limit: "user supplied value or unknown" maximum_spend: "user supplied value or unknown" fees: "value or unknown" decision_deadline: "value or unknown" data_freshness: "timestamp and caveats" risk_review: status: pass | warn | fail | not_run findings: [] blocked_actions: - "order placement, cancellation, routing, signing, and submission" next_safe_step: "one non-executing review action"
End every plan with exactly:
textThis is a planning worksheet, not investment or trading advice. Review venue rules and make any trading decisions yourself.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-05 | pass→pass | 14,461 | 7,925 | -45% | 1 | 1 | 0% | 2,736 | 2,988 | +9% | 0 | 0 | — |
case-01 | fail→pass | 9,643 | 7,813 | -19% | 1 | 1 | 0% | 1,611 | 3,018 | +87% | 0 | 0 | — |
case-02 | fail→pass | 18,224 | 7,902 | -57% | 1 | 1 | 0% | 3,461 | 2,982 | -14% | 0 | 0 | — |
case-03 | fail→pass | 7,041 | 6,537 | -7% | 1 | 1 | 0% | 1,314 | 2,924 | +123% | 0 | 0 | — |
case-04 | fail→pass | 11,893 | 6,610 | -44% | 1 | 1 | 0% | 2,277 | 2,831 | +24% | 0 | 0 | — |
case-06 | pass→pass | 3,479 | 5,879 | +69% | 1 | 1 | 0% | 551 | 2,688 | +388% | 0 | 0 | — |
case-07 | fail→pass | 23,538 | 4,364 | -81% | 1 | 1 | 0% | 3,911 | 2,254 | -42% | 0 | 0 | — |
case-08 | fail→pass | 10,580 | 5,063 | -52% | 1 | 1 | 0% | 1,833 | 2,508 | +37% | 0 | 0 | — |
case-09 | pass→pass | 13,532 | 10,101 | -25% | 1 | 1 | 0% | 2,130 | 3,447 | +62% | 0 | 0 | — |
case-10 | fail→pass | 8,916 | 3,719 | -58% | 1 | 1 | 0% | 1,488 | 2,150 | +44% | 0 | 0 | — |
case-11 | fail→pass | 7,247 | 6,015 | -17% | 1 | 1 | 0% | 1,075 | 2,593 | +141% | 0 | 0 | — |
case-12 | fail→pass | 11,148 | 6,787 | -39% | 1 | 1 | 0% | 1,832 | 2,781 | +52% | 0 | 0 | — |
case-13 | fail→pass | 7,462 | 6,171 | -17% | 1 | 1 | 0% | 1,178 | 2,509 | +113% | 0 | 0 | — |
case-14 | fail→pass | 9,698 | 5,498 | -43% | 1 | 1 | 0% | 1,384 | 2,665 | +93% | 0 | 0 | — |
case-15 | pass→pass | 5,706 | 5,186 | -9% | 1 | 1 | 0% | 911 | 2,555 | +180% | 0 | 0 | — |
case-16 | pass→pass | 12,345 | 7,370 | -40% | 1 | 1 | 0% | 2,103 | 2,928 | +39% | 0 | 0 | — |
case-22 | fail→fail | 12,435 | 5,241 | -58% | 1 | 1 | 0% | 2,137 | 2,522 | +18% | 0 | 0 | — |
case-17 | fail→pass | 8,412 | 2,083 | -75% | 1 | 1 | 0% | 1,466 | 1,898 | +29% | 0 | 0 | — |
case-18 | pass→pass | 9,037 | 4,913 | -46% | 1 | 1 | 0% | 1,787 | 2,508 | +40% | 0 | 0 | — |
case-19 | fail→pass | 9,421 | 5,049 | -46% | 1 | 1 | 0% | 1,583 | 2,483 | +57% | 0 | 0 | — |
case-20 | fail→pass | 11,452 | 5,168 | -55% | 1 | 1 | 0% | 2,130 | 2,596 | +22% | 0 | 0 | — |
case-21 | fail→pass | 10,542 | 5,330 | -49% | 1 | 1 | 0% | 2,151 | 2,644 | +23% | 0 | 0 | — |
case-23 | pass→pass | 11,977 | 1,856 | -85% | 1 | 1 | 0% | 2,033 | 1,843 | -9% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +65 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/27/2026 | +64% |
Other measured skills in the registry, with their headline benchmark lift.