Model Context Protocol server (JSON-RPC 2.0 over plain HTTP POST)

MCP over Streamable HTTP, stateless JSON-response mode. Protocol version is NEGOTIATED at initialize (supported: 2025-11-25, 2025-06-18, 2025-03-26, 2024-11-05 — the server echoes the client's version when supported, else its newest; these are MESSAGE versions — only Streamable HTTP is served, and the deprecated 2024-era dual-endpoint HTTP+SSE transport is NOT provided, so 2024-11-05 is accepted only for clients that can initialize over this endpoint). Each JSON-RPC 2.0 message is its own HTTP POST; responses are plain JSON (no SSE stream, no long-lived connection, no session id issued). Standard lifecycle supported: initialize -> notifications/initialized (HTTP 202, no body; all notifications/* are accepted with 202) -> tools/list -> tools/call; ping also supported. As a documented Signal Bureau CONVENIENCE EXTENSION, direct tools/call without the handshake also works for lightweight stateless integrations. An MCP-Protocol-Version request header, when present, is validated: unsupported values receive HTTP 400. Unknown methods return JSON-RPC error -32601. GET on this endpoint returns HTTP 405 (Allow: POST) with a human-readable discovery body. Tools (29), in the order tools/list serves them, grouped by job: Orient first: start_here; Walk the record: traverse, search_entities, get_entity, get_topic_record, get_events, get_signals, top_accelerating, todays_brief; Trace a claim: ask, get_answer, get_corroborated, get_record, get_calibration; Quote and order: get_quote, create_order, get_order, get_account; Keep a Watch: list_standard_watches, unfurl_watch, save_watch, list_saved_watches, recall_watch, retire_watch, get_watch_feed, get_desk_feed; Feedback and coverage requests: send_feedback, get_feedback_status, propose_topic. Every tool in tools/list carries MCP annotations. Read-only, no side effects (readOnlyHint true): start_here, traverse, search_entities, get_entity, get_topic_record, get_events, get_signals, top_accelerating, todays_brief, get_answer, get_corroborated, get_record, get_calibration, get_order, get_account, list_standard_watches, unfurl_watch, list_saved_watches, get_watch_feed, get_desk_feed, get_feedback_status. Writes state (readOnlyHint false): ask, get_quote, create_order, save_watch, recall_watch, retire_watch, send_feedback, propose_topic; of these, only retire_watch cannot be undone (destructiveHint true). Call start_here first, then walk the record with traverse. To see how the record fits together, start at an entity or event (e.g. Hormuz) and traverse its maintained relationships; each step returns its relation type, receipts, standing and time. traverse opens one node of the record per call: an entity, a situation, a claim or a relation. Each node carries its standing as five separate records (evidence, admission, disposition, validity, measurement), its receipts with source urls, and its neighbours as counts plus typed ids, with a cursor for the next page. One level per call; follow any id it returns to take the next step. Free, no key. The REST twin is GET /api/v1/node/{kind}/{hex}. A first walk in three calls: (1) Open an entity by name. traverse {"id":"entity:hormuz"}, or GET /api/v1/node?id=entity:hormuz: the entity, its standing, its receipts (children), and the situations and claims it is named in (relations). (2) Follow a claim. traverse {"id":"<a relations.items[].id whose kind is claim>"}, or GET /api/v1/node/p/{hex}: the claim in its sources' words, whether it is corroborated, single source or contested, and who carried it. (3) Open its sources. traverse {"id":"<the same claim id>","level":"children"}, or GET /api/v1/node/p/{hex}?level=children: every receipt as an observation with the publisher, the date and the url; open the url to read the source itself. Earlier tool names are PERMANENT aliases, not a deprecation queue — machine callers hold tool names in configs we do not control: orient (now start_here, since 2026-09-07), get_truth_object (now get_topic_record, since 2026-09-07), get_truth_record (now get_topic_record, since 2026-10-01). Each keeps answering, returns the current tool's work, and carries a deprecated marker naming the current name. Earlier prompt names answer the same way: judge-us-first (now look-around, since 2026-10-06), naming the current name in the reply's _meta.alias. An alias never reshapes a payload underneath a caller who never asked for the new name. Prompts (3): morning-read, look-around, coverage-quote. Resources: 8. Rate limits: the ask tool shares the Answer Engine's daily quota and adds a 6/minute burst at the MCP door. Every JSON-RPC error carries a machine-readable error.data.code (parse_error, invalid_request, method_not_found, invalid_params, rate_limited, cannot_measure, data_unavailable, internal_error, or the upstream REST code such as key_required) beside the human sentence in error.message; rate limits add error.data.retryAfter and a Retry-After header. A deploy whose data could not be read answers every method with -32000 / data.state: "cannot-measure" (HTTP 503) rather than an empty result. JSON-RPC batches are refused with -32600.