# Binance Major Crypto Trend > Analyze BTC, ETH, BNB, and other liquid Binance-listed major cryptocurrencies using current macro, spot, derivatives, flows, on-chain, news, event, and X sentiment evidence. Use for trend, up/down probability, buy/hold/reduce/wait posture, relative-strength comparison, or recurring market brief requests; this skill is analysis-only and never places trades. - Skill: `levyflux/binance-major-crypto-trend` (Agent Skill, multi-file: 5 files) - Install (CLI): `npx skillmds@latest add levyflux/binance-major-crypto-trend` - Raw SKILL.md: https://api.skillmd.com/api/skills/levyflux/binance-major-crypto-trend/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: levyflux (https://skillmd.com/u/levyflux) - Updated: 2026-09-22 - Page: https://skillmd.com/skills/levyflux/binance-major-crypto-trend --- # Binance Major Crypto Trend Produce a current, multi-horizon market judgment without turning one indicator, viral post, or headline into a trade call. Default to BTC, ETH, and BNB quoted in USDT when the user does not name assets. This skill is read-only: never log into an exchange, place or cancel an order, sign a transaction, change an account, or claim that monitoring will continue unless the user explicitly requests a supported automation. ## Load the relevant guidance - Read [references/source-playbook.md](references/source-playbook.md) for every current market assessment. It defines source priority, freshness, X collection, and asset-specific evidence. - Read [references/analysis-framework.md](references/analysis-framework.md) when producing an up/down judgment, trading posture, score, scenario, entry/invalidation plan, or cross-asset rank. - Read [references/report-template.md](references/report-template.md) for a full brief or recurring report. A narrow conceptual question does not need the full template. ## Establish the request Infer reasonable defaults instead of blocking: - assets: BTC, ETH, BNB; - venue: Binance spot plus Binance USD-M perpetuals as positioning evidence; - horizons: short 4–24 hours, swing 2–14 days, medium 1–3 months; - quote and timezone: USDT and the user's local timezone, while retaining UTC source timestamps; - output: direction, confidence, conditional posture, triggers, invalidation, and risks. State any meaningful assumption. Ask only when the user wants personalized sizing or an exact plan and account size, current position, maximum acceptable loss, leverage constraints, or horizon would materially change the answer. ## Current-data requirement Browse or query current sources for every live verdict. Record the observation time and source time where available. A remembered price, stale screenshot, search-result snippet, or model knowledge is not current market evidence. If either current core spot price/volume or current derivatives data is missing or stale, do not emit a current short/swing buy/sell or up/down verdict. Give the available structural analysis, label the missing layer, and state what data would unlock a verdict. When X access is partial, continue with a market assessment but label social coverage and lower only the confidence attributable to that layer. ## Evidence contract Classify material claims as: 1. **Verified fact** — current exchange/API data, official release, regulator filing, protocol or chain state, or a directly inspected public statement. 2. **Supported inference** — multiple independent, timely observations favor one explanation. 3. **Weak signal** — plausible but noisy, single-source, socially amplified, or historically unstable. 4. **Unknown** — missing, stale, inaccessible, or unable to distinguish alternatives. Name the evidence layer. Market price, positioning, macro, on-chain, news, and social evidence are not interchangeable. Mirrored articles, reposts, translated copies, and accounts repeating the same origin are one source, not independent confirmation. ## Required analytical posture - Determine the market regime before judging individual assets. - Analyze spot price/volume separately from leveraged positioning. Rising price with rising open interest and positive funding is different from rising price led by spot demand. - Use multiple timeframes; do not let a 15-minute move define a 1–3 month view. - Compare ETH/BTC and BNB/BTC as well as USDT pairs so a nominal rise is not mistaken for relative strength. - Treat X as a narrative, attention, and reflexivity signal—not proof of facts or a standalone trading trigger. Separate credible sources, broad organic discussion, retail emotion, and likely coordinated amplification. - Separate direction from confidence and confidence from data completeness. - Give bull, base, and bear paths with observable triggers. Every actionable posture needs a falsification or invalidation condition. - Never promise returns or present a probability with false precision. Prefer calibrated ranges or qualitative likelihoods tied to evidence. - Do not use universal leverage, stop-loss, take-profit, or allocation values. Derive levels from current structure and volatility. Without a user loss budget, show the sizing formula rather than inventing a position size. ## Decision boundary Translate evidence into both a directional label and a conditional posture: - direction: strong up, up, range/unclear, down, or strong down; - posture: buy-on-confirmation, buy-on-pullback, hold, wait, reduce, exit, or hedge candidate. `Event-risk only` means no new directional position for the affected horizon before the event; wait for post-event spot confirmation plus one independent evidence layer. Apply this veto per horizon: a 4–24 hour event does not automatically veto the 2–14 day or 1–3 month view unless its consequences can plausibly dominate those horizons. Do not issue an unconditional imperative. A posture must specify horizon, trigger, invalidation, and the evidence that would change it. If the user already holds an asset, distinguish hold/reduce/ exit analysis from a fresh entry; ignore the purchase price as a reason to hold. Force the posture to `wait` or `event-risk only` when a decisive claim is unverified, a binary event is imminent, required data is stale, or major evidence layers conflict without a defensible tie breaker. Social enthusiasm alone can never upgrade an asset to a buy. ## Delivery Lead with the timestamped verdict table, then explain the few facts that actually drive it. Include: - market regime and next known risk events; - per-asset direction and confidence for each requested horizon; - conditional posture, trigger, invalidation, and upside/downside scenario levels when supported; - spot-versus-derivatives diagnosis and relative-strength comparison; - macro, flow/on-chain, news, and X findings with confidence labels; - data coverage, freshness, conflicting evidence, and source links; - a concise risk note stating that the result is a conditional market assessment, not guaranteed personalized financial advice. If the user asks for continuous monitoring, define the assets, cadence, alert conditions, timezone, and delivery scope, then use the supported automation mechanism. Do not imply future monitoring from a one-time report.