Nitrogen MCP

Delete a risk questionnaire by ID. Maps to ``DELETE /v1/questionnaires/{questionnaire_id}``. Returns ``{"data": <result>}`` or ``{"error": {...}}``.

questionnaires_delete

Fetch a single check-in by ID. Maps to ``GET /v1/checkins/{checkin_id}``. Returns ``{"data": <check-in>}`` or ``{"error": {...}}``.

checkins_get

Fetch a single client by ID. Maps to ``GET /v1/clients/{client_id}``. Returns ``{"data": <client>}`` or ``{"error": {...}}``.

clients_get

Get the advisor's lead-generation questionnaire embed link key(s). Maps to ``GET /v1/embed_links``. Returns ``{"data": <embed link>}`` or ``{"data": [<embed links>]}`` (or ``{"error": {...}}``). When ``token`` is provided, returns the specific embed link whose key matches that token.

questionnaires_get_lead_generation_key

Fetch a single model or client portfolio by ID. Maps to ``GET /v1/portfolios/{portfolio_id}``. Returns ``{"data": <object>}`` or ``{"error": {...}}``.

models_portfolios_get

Fetch a single retirement map by ID. Maps to ``GET /v2/retirement/{retirement_id}``. Returns ``{"data": <retirement map>}`` or ``{"error": {...}}``.

retirement_maps_get

Fetch a single risk questionnaire by ID. Maps to ``GET /v1/questionnaires/{questionnaire_id}``. Returns ``{"data": <questionnaire>}`` or ``{"error": {...}}``.

questionnaires_get

List the advisors you may act on behalf of. Maps to ``GET /v1/advisor/act_as_options``. Call this before passing ``act_as_advisor_id`` to a tool that accepts it — only some tools do, and a tool accepts it only if ``act_as_advisor_id`` appears in its own input schema; passing it to any other tool is rejected before the call is made. That argument takes an ``rsk_user_id``, which is an internal identifier — it cannot be guessed, and a wrong value is refused with a message that deliberately reveals nothing about which advisors exist. This is the only way to obtain a usable one. Each entry carries ``rsk_user_id`` plus the advisor's name and email, so a person can be shown which advisor they are choosing rather than a number. Returns an empty list if you are not an assistant for anyone — that is a success, not an error. One response carries at most 100 advisors. When more than that match, the response says so in ``truncated`` and names the ``next_offset`` to ask for: every advisor you may act as is reachable by paging, or directly by ``search``. Never treat a page as the whole list.

advisors_get_act_as_options

List check-ins for the authenticated advisor. Maps to ``GET /v1/checkins``. When ``client_id`` is provided, returns check-ins for that client. Otherwise returns the advisor's full check-in list. Returns ``{"data": <check-ins>}`` or ``{"error": {...}}``. The Core API serves those two modes from different code paths, so some filters are mode-specific: ``only_latest`` requires ``client_id``, while ``show_all``, ``exclude_analytics``, and ``limit`` only apply to the full advisor list (omit ``client_id``). Using a filter with the wrong mode returns a validation error instead of silently ignoring it.

checkins_get_list

List the authenticated advisor's clients. Maps to ``GET /v1/clients``. Returns ``{"data": [...]}`` by default, or ``{"error": {...}}`` on failure. Set ``no_share=True`` to return only clients the advisor owns.

clients_get_list

List the authenticated advisor's models. Maps to ``GET /v1/portfolios``. Returns ``{"data": [...]}`` or ``{"error": {...}}``. The advisor is identified by the session; no parameters are required.

models_portfolios_get_list

Search for clients by name or other criteria. Use this tool to look up client identities before passing a client_id to the nucleus_agent tool.

client_search

Search clients by name only. Maps to ``GET /v1/clients/search?q=...``. Returns ``{"data": [...]}`` or ``{"error": {...}}``.

clients_search_by_name

Search clients by name or email. Maps to ``GET /v1/clients?q=...``. Returns ``{"data": [...]}`` or ``{"error": {...}}``.

clients_search_by_name_email

Set a preferred proxy security for an allocation in a portfolio. Maps to ``PUT /v1/portfolios/{portfolio_id}``, setting the target allocation's ``adjustments.proxy.value`` to the Riskalyze Security ID. This is a read-modify-write: it fetches the portfolio, injects the proxy into the named allocation, and writes the whole portfolio back so nothing else is lost. Returns ``{"data": <updated portfolio>}`` or ``{"error": {...}}``.

models_portfolios_update_preferred_proxy

Sync a client with their integrations. Maps to ``GET /v1/integration/sync_client_generic/{client_id}``. Triggers a sync for the client across linked integrations. Returns ``{"data": <result>}`` or ``{"error": {...}}``.

clients_sync_with_integrations

Update an existing client. Maps to ``PUT /v1/clients/{client_id}``. Returns ``{"data": <updated client>}`` or ``{"error": {...}}``.

clients_update

Update an existing account in a model or client portfolio. Maps to ``PUT /v1/portfolios/{portfolio_id}``. Returns ``{"data": <updated portfolio>}`` or ``{"error": {...}}``. The ``account`` MUST contain its ``id``. The account is replaced wholesale, so pass the complete account (all allocations to keep). This is a read-modify-write: the current portfolio is fetched and written back with only the matching account replaced, so all other accounts and the portfolio's ``client_id``/``name``/``type`` are preserved automatically — you do not pass them.

models_portfolios_update_account

Update an existing retirement map. Fetches the current map, merges the provided settings into the existing ``settings`` object, and PUTs the full merged ``settings`` to ``/v2/retirement/{retirement_id}``. Provided fields overwrite existing fields; omitted fields are preserved. Returns ``{"data": <updated retirement map>}`` or ``{"error": {...}}``.

retirement_maps_update