Energy Aspects

**MANDATORY FIRST STEP — do not skip: before your first call to this tool in this conversation, load skill://ea_mcp_skill and schema://ea-live-posts — either read them as MCP resources, or (if your client cannot read MCP resources) call the `get_skill` tool once per URI.** They define author-attribution rules (no individual analyst names) and topic-tag caveats this tool's results depend on. Retrieve EA Live posts — short, real-time market updates and analyst commentary — via semantic search. Prefer over `get_publications` for recent news or intraday commentary. Parameters: `query` (required, str) — natural-language or keyword description; `from_` (optional, str) — earliest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be a valid calendar date); `to` (optional, str) — latest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be valid); `limit` (optional, 1–20, default 10) — number of results to return. Response is XML-formatted with <context> wrapper. Returns a XML document with results containing: title, metadata_content, published_at, content_type, chunks, licenses, url. See schema://ea-live-posts resource for detailed field definitions. For questions spanning multiple sources, prefer `get_insight_and_data`. IMPORTANT: Always cite sources using the `url` field as a markdown link: [Title](url). VALIDATION: from_, to, and limit must pass server-side validation or the tool returns a validation_error.

get_ea_live_posts

**MANDATORY FIRST STEP — do not skip: before your first call to this tool in this conversation, load skill://ea_mcp_skill and the relevant schemas (schema://publications, schema://ea-live-posts, schema://timeseries) — either read them as MCP resources, or (if your client cannot read MCP resources) call the `get_skill` tool once per URI.** Retrieve publications, EA Live posts, and timeseries data in a single parallel request. This is the preferred starting tool for broad or initial research questions — use the dedicated tools (`get_publications`, `get_ea_live_posts`, `get_timeseries_data`) only when targeting a specific content type. Read-only. Parameters: `query` (required, str) — natural-language or keyword description of the topic; `from_` (optional, str) — earliest date filter applied to all content types (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be a valid calendar date); `to` (optional, str) — latest date filter (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be valid); `limit` (optional, integer 1–20, default 10) — maximum results returned per content type. Response is XML-formatted with <context> wrapper. Returns data as XML document with keys `publications`, `ea_live`, and `timeseries` (each containing results with the same schema as their respective dedicated tools) plus an `errors` object reporting any per-source failures — partial results from successful sources remain valid even when one source errors. IMPORTANT: Always cite sources using the `url` field formatted as a markdown link: [Title](url). VALIDATION: from_, to, and limit must pass server-side validation or the tool returns a validation_error.

get_insight_and_data

**MANDATORY FIRST STEP — do not skip: before your first call to this tool in this conversation, load skill://ea_mcp_skill and schema://publications — either read them as MCP resources, or (if your client cannot read MCP resources) call the `get_skill` tool once per URI.** They define citation, provenance, and vintage-disclosure rules this tool's results depend on. Retrieve Energy Aspects in-depth research publications via semantic search. Parameters: `query` (required, str) — natural-language or keyword description; `from_` (optional, str) — earliest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be a valid calendar date); `to` (optional, str) — latest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be valid); `limit` (optional, 1–20, default 10) — number of results to return. Response is XML-formatted with <context> wrapper. Returns a XML document with results containing: title, metadata_content, published_at, content_type, chunks, licenses, url. See schema://publications resource for detailed field definitions. For questions spanning publications, EA Live posts, and timeseries, prefer `get_insight_and_data`. IMPORTANT: Always cite sources using the `url` field as a markdown link: [Title](url). VALIDATION: from_, to, and limit must pass server-side validation or the tool returns a validation_error.

get_publications

Return the text of an EA MCP skill or schema document. Use this when your client cannot read MCP resources, to satisfy the MANDATORY FIRST STEP of the other tools; it returns exactly the same content as reading the MCP resource with that URI. Read-only. Parameter: `uri` (required, str) — one of: skill://ea_mcp_skill, schema://publications, schema://ea-live-posts, schema://timeseries. Returns the document as plain text (markdown for skill:// URIs, JSON for schema:// URIs), or an error message listing the valid URIs.

get_skill

**MANDATORY FIRST STEP — do not skip: before your first call to this tool in this conversation, load skill://ea_mcp_skill and schema://timeseries — either read them as MCP resources, or (if your client cannot read MCP resources) call the `get_skill` tool once per URI.** The skill includes the series confirmation gate, which MUST be passed before calling this tool — do not call it on an unconfirmed series. Retrieve Energy Aspects time-series datasets — historical values and forecasts — via semantic search. Parameters: `query` (required, str) — natural-language description (e.g., 'US crude oil production', 'European natural gas demand'); `from_` (optional, str) — earliest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be a valid calendar date); `to` (optional, str) — latest date (yyyy-MM-dd or yyyy-MM-ddTHH:mm:ssZ, must be valid); `limit` (optional, 1–20, default 10) — number of datasets to return; `release_date` (optional, yyyy-MM-dd, must be a valid calendar date) — point-in-time query (data as it appeared on that date). Response is XML-formatted with <context> wrapper. Returns a XML document with results containing: metadata (with description, country, region, aspect, unit, frequency, lifecycle_stage), data object (date → numeric value mapping), dataset_id, url. See schema://timeseries resource for detailed field definitions. For questions spanning multiple sources, prefer `get_insight_and_data`. IMPORTANT: Always cite sources using the `url` field as a markdown link: [Title](url). VALIDATION: from_, to, release_date, and limit must pass server-side validation or the tool returns a validation_error.

get_timeseries_data