Loading ToolBench
Optic reads every market that prices the same story — the Binance exchange, onchain, prediction markets and social attention — and reports where they stop agreeing. Cross-venue market analysis and discovery engine.
Optic has 11 tools with clear, domain-specific descriptions (avg 180 chars, within baseline 194). Tool names use domain verbs (optic_read, optic_scan, optic_rug) but lack standard verb_noun patterns (get_, list_, search_). Input schemas are present for most tools but inconsistent: optic_read and optic_rug have detailed Zod schemas with constraints; optic_scan, optic_daily, optic_edge, optic_smart_money, optic_pulse, optic_status have empty input objects ({}). Output schemas are not documented, responses are returned as structuredContent but the shape is not declared. Error handling is absent; no recovery guidance or error classification. Parameters lack descriptions in several tools (optic_scan, optic_daily, optic_edge, optic_smart_money, optic_pulse, optic_status have no params to describe, but this is a design choice). optic_ticket has a WRITE risk classification but no confirmation/dry-run pattern. Overall: solid descriptions and naming intent, but schema inconsistency and missing output documentation hold the score back.
Today's strongest signals across prediction markets and onchain momentum, as decisive research-backed calls with the evidence. Confidence is signal strength, never a claimed win rate.
Where a prediction market's priced probability looks soft or rich against live web research. Conservative by design — the market is usually right, and an honest 'nothing today' is a valid answer.
BTC, ETH and SOL short-horizon Up/Down windows on Binance Wallet prediction markets, set against the live Binance spot tape — where the market leans versus where price actually is.
The full read on any coin, token, contract address, event or story: the Binance exchange (spot + perps positioning), onchain data, Binance Wallet prediction markets and the social-hype board, compared in one call. Returns a 0-100 gap score for how far apart those markets are, the reasoning with the numbers behind it, and a shareable card. Data and analysis only.
Token security audit plus holder composition into a 0-100 risk score with the concrete red flags: contract mechanisms, transfer taxes, dev/sniper/bundler/insider concentration. Risk disclosure, never advice.
Six tools (optic_scan, optic_daily, optic_edge, optic_smart_money, optic_pulse, optic_status) have empty input schemas ({}). While intentional for discovery tools, output schemas are not documented anywhere. LLMs cannot plan downstream calls or extract structured data without knowing what fields to expect.
Tool names do not follow standard verb_noun convention (get_, list_, search_, create_). Names like 'optic_read' and 'optic_scan' are domain-specific but lack the action-verb clarity that helps LLMs parse intent. Baseline: 90% of A+ tools start with action verbs.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 71 | 2025-06-18+ | v2 |
Discovery over the whole market: which tokens' social hype is accelerating before the crowd, which AI-detected narratives are pulling inflow, what just launched with real activity, and where smart-money inflow concentrates.
Tokens that Binance-tracked smart-money wallets are accumulating right now, by wallet count and inflow. Factual flow, never a trade instruction.
Which exchange kit this Optic deployment runs on, whether it holds its own Binance MCP session, and the HTTP base URL of the same services.
One company across markets: the exchange-listed tokenized share price on-chain, the real-world close and analyst consensus from live research, and any prediction market on the company.
Turns a decision the CALLER has already made into a ready-to-place order on a Binance Wallet prediction market: resolves the live market, reads the priced outcome, sizes the position and returns the exact `baw prediction trade quote` and `place-order` commands for the caller's own Binance Agentic Wallet. Optic never chooses the side and never signs — run the quote, show it to the user, and only place after they confirm.
Lifecycle stage for a token — igniting, building, peaking, cooling or quiet — from its age, its social-hype trajectory and its onchain activity acceleration.
optic_ticket is marked WRITE risk but has no confirmation/dry-run pattern. The tool returns a quote and commands for the caller to execute, which is good, but the description does not explicitly state that the caller must confirm before placing the order. This should be reinforced in the description.
No error handling or recovery guidance. Tools do not document what errors can occur, how to classify them (retryable vs user-fixable), or what the LLM should do next. Baseline: 100% of A+ tools have error classification.
optic_read accepts 'binance_market_data' as type 'unknown' with a long description but no schema validation. The description explains the expected format (object, array, or JSON string) but the Zod schema does not enforce it. This invites malformed input.