UNIFIELD

Read-only advisory/safety check. Tests whether a proposed agent action is consistent with a stored intent's core, axiom exclusions, and modifiers. Returns allowed | needs_confirmation | blocked. An ok:false / 'blocked' result MUST be treated as a stop-or-clarify signal — do not override it on the user's behalf. Advisory only — humans remain responsible.

Read-only. Explain a match between two intents in two modes. Non-binding: humans remain responsible for any action. - mode='evidence' (DEFAULT for MCP): returns canonical structured evidence with NO server LLM call and NO premium_explanations charge. REQUIRES match_id. Suitable for ChatGPT App / agentic clients which formulate the natural-language explanation locally and EPHEMERALLY — UNIFIELD never persists ChatGPT-generated wording. ChatGPT may explain the evidence to the user, but MUST NOT invent evidence, alter scores/grades, or expose counterpart original text. reasoning_source remains 'algorithmic' | 'llm' (no 'chatgpt_app' value is introduced). - mode='narrative': returns server-generated NL prose (legacy default for REST). Charges premium_explanations on cache miss; cached explanations (when match_id is supplied and explanation_input_hash matches) do NOT charge. Original-text boundary (6B-3, no migration): • intent_original_counterpart is ALWAYS null (counterpart raw original_text is reserved for 6C share_original_text consent). • intent_original_self is returned only when include_original_text=true AND the caller owns source_intent_id AND the stored intent has a confirmed original (original_text_mode='store_private'). • Chat transcripts are NEVER acceptable as original_text. ChatGPT-originated intents must field a single user-confirmed formulation, never raw chat history.

Read-only. Find possible complementary intents for a user-owned intent across the Unifield (and any private fields the caller has access to). Use this to surface candidates for the user; results are non-binding suggestions, never guarantees, and human judgement is required to act on them. Filters: limit, min_score, field_id, kind, status. Owner-only refresh. include_explanations=true generates structured explanations for the top max_explanations matches (absolute v1 ceiling 3; further reduced by per-plan billing_plans.monthly_quota.max_explanations_per_request, default 1 free / 3 paid). The response surfaces filters.max_explanations_requested, filters.max_explanations_effective, and filters.plan_cap_source (request | plan | ceiling). Stored explanations are cached with explanation_input_hash; on each call the server recomputes the hash. Match.explanation_status: stored (hash matches) | stored_unknown_hash (legacy explanation, no hash on file) | stale (hash differs, explanation_error=input_changed) | generated | skipped | failed. refresh_explanations=true regenerates stale or hash-unknown explanations. explanation_error values: input_changed | plan_cap | beyond_max_explanations | quota_exceeded | rate_limited | ai_error | ai_unreachable | not_authorized. Stored/stale/stored_unknown_hash do not bill premium quota; only generation attempts (success or failure) bill once. matching_status is computed (queued | processing | completed | no_matches | failed). No paid ranking — plans only cap explanation depth, never score, ordering, or candidate inclusion.

Read-only. Return the fielding lifecycle and stepper status of a user-owned intent (draft → analysis → active → connected). Use this to help the user see where their intent stands in the Unifield without modifying anything.

Read-only. List the fields the user may field intents into. Distinguishes the open global Unifield from private fields provided by UNIFIELD (e.g. organisations, communities). Use this to help the user choose where to field an intent.

Preview a possible intent (IXP intent molecule) from user-provided wording. Read-only: nothing is stored or fielded by this call. Use this to help the user articulate what they want before fielding it into the Unifield. For ChatGPT-originated flows, do NOT pass arbitrary chat history as `text`; ask the user to confirm a single editable formulation first. Chat context is never original_text. Output is a non-binding preview; the user must confirm before publish_intent fields anything.

Field a user-confirmed intent into the Unifield (the open global intent layer) or into a private field provided by UNIFIELD. Mutating. Private by default. Use this to help the user place a confirmed formulation into the Unifield AFTER you have shown the proposed intent and obtained explicit human confirmation in the current turn. Literal `consent_confirmed:true` is REQUIRED whenever the target is the global Unifield, `global_unifield:true`, or visibility/consent_state is non-private; missing/false consent returns JSON-RPC -32013 consent_required. Original-text boundary: `original_text_sharing:"omit"` is the safest default for ChatGPT-originated flows; `"store_private"` is only acceptable when the user has explicitly confirmed a single editable formulation. Raw chat context must NEVER be sent as original text. Do not phrase to the user as 'publish' — say 'field' / 'into the Unifield'. Non-binding: humans remain responsible.

Render a single complement match card from find_complements.matches as a widget surface. Read-only: never requests contact, never calls request_connection, never calls any backend or tool. Input must be one privacy-safe ComplementMatchCard. Never exposes counterpart original_text or chat context.

Render a previously parsed intent draft (IntentDraftPreview) as a widget surface. Read-only: performs no writes, makes no backend or other tool calls, never re-fetches data. Input must be a strict subset of parse_intent's structuredContent envelope. Never accepts or exposes counterpart original_text or chat context.

Render the lifecycle stepper for a caller-owned intent (from get_intent_status or publish_intent). Read-only: performs no writes, makes no backend or other tool calls. Never exposes counterpart original_text or chat context. Either id or intent_id, and either status or current_step, must be provided.