> For the complete documentation index, see [llms.txt](https://nytshift.gitbook.io/nytshift-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://nytshift.gitbook.io/nytshift-docs/agentops/current-capabilities.md). # Current capabilities [← AgentOps overview](/nytshift-docs/agentops/agentops.md) · [Documentation home](/nytshift-docs/start-here/readme.md) · Next: [Qualification loop →](/nytshift-docs/agentops/qualification-loop.md) This page is the public claim boundary for the AgentOps narrative. Product availability still depends on the exact route, venue, account scope, evidence freshness and operator configuration reported by the interface. ## Capability ledger ### Market context · CURRENT Available where reported. Source-labelled Arcus, Lighter and Hyperliquid data can appear in the operating picture. Unsupported, missing or stale inputs fail visibly. ### Robinhood Chain intelligence · CURRENT Read-only identity, provider and chain evidence remain separate from venue-order authority. ### Human PAPER · CURRENT Available browser-local deterministic simulation uses fresh normalized book evidence and creates no venue order. ### Human trading workflows · GATED The terminal can construct and review exact venue actions, but submission depends on that venue's independent signer, readiness, reconciliation and authorization state. ### Agent PAPER · GATED Local-only. A valid signed Agent Pro entitlement and NYTSHIFT platform-operator-issued connection are required. Hosted customer self-service MCP is not current. ### Agent market and account research · GATED Entitled, issued tools enforce exact symbols and account grants. Public addresses still require an explicit grant before external-model use. ### Agent proposals · GATED Advisory only. An entitled, issued credential can register and track a strict proposal, but does not reserve or grant signer authority. ### Agent activity · GATED Partial foundation. An issued credential can read its bounded activity; a complete cross-surface Flight Recorder is not. ### AEGIS · CURRENT Deterministic review currently resolves approved or rejected decisions from server-owned evidence. ### Hyperliquid testnet policy · GATED A separately versioned automation path requires the signed Agent Pro v2 automation add-on, an active policy and its own sender-bound request proof. It may make one bounded signer attempt per request identity and consumes the add-on's monthly automation quota. ### Mainnet agent authority · UNAVAILABLE No MCP or automation mode exists. Architecture and readiness documents do not constitute authority. ## What is next * one durable `PaperRun` model across human and agent experiments; * a Flight Recorder binding context, tool activity, proposal, policy and PAPER outcome to an exact strategy version; * a private strategy registry with version history, lineage and comparison; * a Strategy Passport containing mandate, assumptions, evidence windows and unresolved limits; * a broader AEGIS receipt vocabulary for `ALLOW`, `RESIZE`, `HOLD` and `BLOCK`; * hosted tenant-safe agent access only after its OAuth, TLS, revocation, abuse-control and security evidence is complete; * venue and account views organized by human, agent and strategy attribution without hiding partial data. ## What is vision * broad all-perpetual-venue coverage; * automatic best-venue routing; * cross-venue live execution under one generic agent permission; * production-qualified mainnet agent automation; * public strategy rankings, competitions, licensing or sales; * autonomous promotion from PAPER to live capital. Vision items are not release promises. Any future capital-moving mode must retain venue-specific signer, eligibility, reconciliation, alerting and revocation evidence. ## Non-negotiable boundaries * An MCP connection is not a wallet or signer proxy. * A profitable PAPER run does not grant a wider mode. * Payment or plan state does not grant execution authority. * A signed Agent Pro entitlement enables bounded credential issuance; it does not prove payment or grant signer authority. * An agent cannot change its own policy, scope, quota or expiry. * A stale, partial or unknown state never becomes approval by omission. * A changed strategy version does not inherit qualification automatically. * NYTSHIFT does not train an external model's weights. Agents and humans iterate on strategies inside bounded experiments. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://nytshift.gitbook.io/nytshift-docs/agentops/current-capabilities.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.