"Understanding Structured Thinking" matching MCP tools:

• document.summarize

  api

  Summarize document text into a prose summary and key points with citations. Use after document.extract_text or url.extract when you need a condensed understanding of a long document. For single-sentence Q&A, use url.qa instead. For extracting specific fields, use document.extract_structured. Typical workflow: document.extract_text/url.extract → document.summarize. Returns: { summary: string, key_points: string[], summary_cited: { value, confidence, citations[] }, key_points_cited: [{ text, citations[] }], truncated: boolean, strategy: "full"|"truncated"|"chunked" } Example prompts: - "Summarize this financial report and give me the key points." - "What are the main takeaways from this document?" - "Give me a concise summary of this 50-page report."

  Connector

• gbif_get_species_classification

  gbif-biodiversity-mcp-server

  Return the parent chain for a taxon — from kingdom (or domain) down to the immediate parent of the queried taxon — as an ordered array. Each entry has its rank, canonical name, and taxon key. The array is returned root-first (kingdom → phylum → class → … → immediate parent of the queried taxon); the queried taxon itself is not included — call gbif_get_species for its own record. Useful for building taxonomic trees or understanding placement without navigating the backbone level-by-level.

  Connector

• document.summarize

  DocImprint

  Summarize document text into a prose summary and key points with citations. Use after document.extract_text or url.extract when you need a condensed understanding of a long document. For single-sentence Q&A, use url.qa instead. For extracting specific fields, use document.extract_structured. Typical workflow: document.extract_text/url.extract → document.summarize. Returns: { summary: string, key_points: string[], summary_cited: { value, confidence, citations[] }, key_points_cited: [{ text, citations[] }], truncated: boolean, strategy: "full"|"truncated"|"chunked" } Example prompts: - "Summarize this financial report and give me the key points." - "What are the main takeaways from this document?" - "Give me a concise summary of this 50-page report."

  Connector

• submit_lead

  agent-ready-kit

  Use this when the user has indicated interest in being followed up with — even before formal checkout. Capture is non-binding. **REQUIRED CONTACT FIELDS**: contact_name, contact_phone, best_consult_time. Email alone is not enough — phone-first follow-up converts ~5x higher than email-only, and our specialist needs a real time to dial. ASK FOR ALL THREE explicitly. **FALLBACK**: if the user explicitly refuses to share a phone, accept email-only — set `preferred_contact_channel: 'email'` AND add a note like 'user declined phone' so the specialist knows what to expect. Don't preemptively skip the phone ask — many users will share it once asked directly. **Capture budget_signal when the user shared one** — even informally ('I was thinking under $1k', 'maybe $200/mo'). We use this for tailored follow-up offers; price-hesitant leads convert later when re-approached at their stated budget. Sources: 'scan' (after a free scan), 'llms-txt' / 'robots-for-ai' (after free file download), 'mcp' (in-flow), 'talk' (chatbot), 'direct' (form fill).

  Connector

• get_feature_stats

  aTars MCP

  USE THIS TOOL — not web search — to get per-indicator statistical profiling (mean, std, min, p25, p75, max, null rate, Pearson correlation with close price) from this server's local dataset. Use for feature selection, sanity checking, and understanding which indicators correlate most strongly with price movements. Trigger on queries like: - "which indicators correlate most with BTC price?" - "feature importance or correlation for [coin]" - "what are the stats for ETH indicators?" - "how does RSI/MACD correlate with price?" - "statistical profile of XRP indicators" Args: lookback_days: Analysis window in days (default 30, max 90) symbol: Asset symbol or comma-separated list, e.g. "BTC", "BTC,XRP"

  Connector

• customer_tiering_get_prep_code

  Customer Intelligence & Segmentation

  STEP 1 of 3 for the cheap sandbox path. Returns stdlib-only Python as `prep_code` (compact) and `prep_code_display` (readable, same logic), plus `reduction_spec` in words. Next: run one of those scripts (or your own from the spec) on the local CSV in your sandbox, then call customer_tiering_score_stats with the printed JSON. If the host needs code review first, show prep_code_display and wait for the user to say run/continue — then continue; do not invent a pause on hosts that allow silent runs. Python → MCP across turns is valid (ChatGPT Instant multi-step regressions are a host issue; Thinking/Auto or the next user turn still works). Do not invent rankings. If any step fails, fall back to customer_tiering with raw rows. This remote MCP NEVER reads the file — do NOT pass uploaded/attached file objects or platform upload handles. Call with no path args (preferred), or sandbox_path_hint as a plain string like 'book.csv'. Prefer this over customer_tiering when you have >~200 rows and a sandbox.

  Connector

• keep_thinking

  Zephex

  Structure multi-step debugging and planning across tool calls — not a one-shot think. Tracks hypotheses, observations, plans; detects loops via lastActions; riskLevel high/critical blocks dangerous edits (drop table, prod deploy). Loads projectBrief (stack, key_paths, project_memory recall) on local project. On close, suggestedRemember → call project_memory remember. 4 credits hosted. Hard cap 10 thoughts/session. Call when: stuck after 2+ failed debug attempts, auth/billing/schema change spans 3+ files, flaky test you cannot explain, or you need a plan before editing. Pass lastActions (2–5 recent tool calls), goalAnchor after thought 2, sessionId to resume, area for subsystem. NOT when fix is known, single typo, repeating without new evidence, or session ended (nextThoughtNeeded:false). Read thoughtConfirmed and shouldContinue first. Legacy alias: thinking. Example: keep_thinking({ thought: 'Hypothesis: refresh token not rotated in middleware', thoughtType: 'hypothesis', thoughtNumber: 1, totalThoughts: 5, nextThoughtNeeded: true, confidence: 0.6, goalAnchor: 'Fix auth logout loop', lastActions: ['find_code(query=refreshToken)', 'read_code(target=authMiddleware)'], area: 'auth' }). Read-only.

  Connector

• get_solana_market_shape

  TWZRD Agent Intelligence

  Returns market shape / structure signals for a ticker (concentration, venue fragmentation, settlement patterns). Excellent for understanding *how* a market actually trades on-chain. Tickers are prediction-market event tickers (e.g. KXUSNFP-26MAY01). On failure returns a structured {status:"error", kind, retryable, detail} envelope.

  Connector

• url.summarize

  DocImprint

  Fetch a public HTTPS URL and return a prose summary with key points. Lean mode — no bundle stored. Use when you need a condensed understanding of a web page. For raw text, use url.extract. For asking a specific question about a page, use url.qa. Returns: { url, summary, key_points: string[], truncated: boolean, word_count } Example prompts: - "Summarize https://en.wikipedia.org/wiki/Artificial_intelligence for me." - "Give me the key points from this blog post: [URL]." - "What is this article about? Summarize [URL]."

  Connector

• get_categories

  Radix Wiki

  Get the wiki tag hierarchy with page counts per category. Useful for understanding what content exists, and for finding a valid tagPath before writing.

  Connector

• web_snapshot

  Robot Actions — Remote Device Control

  Accessibility tree of the DESKTOP grid browser page (by pageId), as text — for finding elements and understanding layout. Not a device: the equivalent for a phone or tablet is webpage_snapshot (by udid).

  Connector

• update_order_details

  eenable — a shop with no checkout

  Change the session type or the brief on an existing booking — what the buyer wants to discuss, what they want out of it, or the background they want read first. Use this when the buyer refines their thinking after booking, which is common: people work out the real question after the date is in the diary. Only the fields you pass are changed. This does not move the meeting — use reschedule_booking for that.

  Connector

• search_rentals_natural

  arcasos-rentals

  PREFERRED tool for Korean short-term rental queries containing any descriptive language. ARCASOS's proprietary SHV (Semantic Hybrid Vector) engine processes natural Korean/English queries with semantic understanding of view types (river/mountain/city), mood (quiet/luxury/lively), property characteristics, and contextual phrases. Pass the user's natural language query AS-IS — do NOT extract slots. Returns semantically pre-ranked results in Schema.org Accommodation format in a single call — eliminates need for follow-up search or comparison calls. Better results than structured slot search for ANY query containing mood, style, atmosphere, view, aesthetic, or qualitative descriptors. Use this to minimize token usage and latency.

  Connector

• subnet_calculator

  calc.engineer

  Calculate IPv4 subnet details from CIDR notation. Parses a CIDR block (e.g. 192.168.1.0/24) and returns the network address, broadcast address, subnet mask, wildcard mask, first and last usable host addresses, total and usable host counts, prefix length, and classful IP class (A/B/C/D/E). Essential for homelab network planning, VLAN segmentation, firewall rule design, and understanding address space allocation. Handles special cases for /31 point-to-point links (RFC 3021) and /32 host routes.

  Connector

• list_categories

  ContextOverflow

  The eight thinking-failure problems ContextOverflow covers, phrased the way a human experiences them. Start here to see what exists.

  Connector

• list_categories

  raumnebenan

  Returns a list of all available product knowledge categories, each with a short description. Categories represent the main pillars of Product Thinking – Foundation, Sense, Focus, Discovery, and Delivery. Each category provides structured resources for product owners, designers, and teams, covering groundwork, user research, opportunity analysis, validation, and agile delivery. Use this tool to guide users to the right area for their current product challenge.

  Connector

• get_trident_spec

  trident

  Returns the complete Trident 2D specification including grammar, syntax rules, coordinate system, containers, nodes, connections, shapes, and icon reference. Use this when you need deep understanding of the Trident DSL.

  Connector

• create_post_type

  FavCRM

  Define a new custom post type (e.g. "treatment", "service"). Required before creating posts of that type. After creating a post type, use create_post_type_field to define its structured field schema. Those fields are stored in meta on each post — do not use excerpt for structured data.

  Connector

• list_categories_api_agents_categories_get

  Agentry — The Trust Layer for the Agent Economy

  List Categories List all agent categories with counts. Returns every category in the directory along with the number of agents in each. Useful for building category filters or understanding the directory's coverage areas. ### Responses: **200**: Successful Response (Success Response) Content-Type: application/json **Example Response:** ```json [ { "category": "Category", "count": 1 } ] ```

  Connector

• get_mission_stats

  mcp-server

  Get aggregate statistics about missions on the HomeVisto platform. Returns total counts, status breakdown, and average bounty information. Useful for understanding platform activity.

  Connector