"Red Hat" matching MCP tools:

• get_losing_positions

  OneQAZ Trading Intelligence

  Purpose: Losing paper positions (ROI < 0). Convenience wrapper around get_positions(max_roi=-0.01). Triggers (casual questions too): "what's underwater?", "지금 뭐가 물려 있어?", "show me the red ones", "any positions in trouble?", "얼마나 손실 중이야?". When to call: drawdown / risk review. Prerequisites: none. Next steps: get_position_detail, get_role_analysis. Caveats: paper-trading data only. Args: market_id: Market ID (crypto, kr_stock, us_stock; aliases coin/kr/us accepted) limit: Max results (default 20) Disclaimer: Information only, not investment advice.

  Connector

• picsart_preflight

  Picsart GenAI

  Free pre-flight check before `picsart_generate`: in ONE call it (1) validates a candidate params object against the model's parameter schema + inter-parameter constraints, and (2) quotes the credit cost — without running the model or charging the user. Use this after assembling params (user input, derived defaults, model swaps) and before generating, to surface bad arguments and show cost. Do NOT use it to look up which params a model accepts (use `picsart_model_params`) or to actually generate (use `picsart_generate`). Required inputs: `model` id and a `params` object (put the `prompt` inside `params`). Example: `{ model: "flux-2-pro", params: { prompt: "a cat in a hat", aspectRatio: "1:1", count: 1 } }`. Returns `{ model, valid, errors?, credits }`: `valid`/`errors` are from local validation (always present, no auth needed; `errors` only when invalid); `credits` is the dry-run cost (a number), or `null` when pricing is unavailable or the request is unauthenticated.

  Connector

• verify_payee

  Imperio — Italian Tax & Compliance Tools

  Verifica anti-frode di un beneficiario PRIMA di pagare una fattura (contro la truffa del cambio-IBAN / Business Email Compromise). Controlli: IBAN (checksum MOD-97 + banca dal codice ABI), Partita IVA e Codice Fiscale (checksum), e — quando il servizio UE è raggiungibile — la ragione sociale reale via VIES live con match sul nome atteso. Restituisce un verdetto (ok / attenzione / alto_rischio) + i red flag puntuali. A differenza degli altri strumenti fa I/O esterno (VIES): se VIES è giù o il budget di verifica è temporaneamente esaurito, il verdetto degrada ONESTAMENTE ad "attenzione" (mai un falso "ok"). Gratis (€0), nessun login. NON sostituisce la verifica out-of-band (telefonata al fornitore su un numero già noto).

  Connector

• brc_start_company_connection

  Red by Big Red Cloud

  Starts the secure Red company connection flow and generates a fresh one-time secure Red connection link and confirmation code. Use only when there is no active company connection, no valid connectionRef, the user explicitly asks to connect or reconnect, try again after a failed connection, expired session credentials, or when an old, used, or stale secure connection link no longer works. Do not call this tool when a valid connectionRef from brc_confirm_company_connection is already available and recent tool calls succeeded with it. Do not call this tool because a lookup returned no rows, partial data, or an empty list — that means no matching records, not an expired connection. Do not call this tool after successful company data retrieval unless the user explicitly asks to connect, reconnect, or add more companies. Always call this tool again to generate a new link — never reuse a previous connection link. Returns a one-time connection page URL (no time expiry, but each link works only once). On that page the user can enter a single company or upload a CSV for multiple companies — never in chat. After completing the secure page, the user should return to this chat and provide (copy/paste) the confirmation code shown on the success page. Do not paste an API key into chat.

  Connector

• brc_send_sales_invoice_email

  Red by Big Red Cloud

  Sends a sales invoice email. Supported document type only. Red email sending is available for sales invoices, quotes, and customer statements — not for cash receipts, purchases, payments, bank accounts, customers, suppliers, products, reports, or other document types. If the user asks to email an unsupported document type, say Red cannot email it through the current MCP tools, list the supported types, and stop without preparing an email preview or attempting a workaround. Do not call this tool with confirmSend=true until the user has reviewed a plain-English email preview and explicitly confirmed they want to send it. The email preview must show the recipient email address clearly before asking for send confirmation. If there is no customer email on file and no recipient override, stop and ask for a recipient email address — do not send. Create/post confirmation and email send confirmation are separate steps. If the user provides multiple recipient addresses, ask whether to send one email using BCC or separate individual emails. Only use sendMode='separate' when the user explicitly chooses separate emails. Do not ask about BCC unless the user provides multiple recipients or asks to copy another address. Requires routeToken from brc_route_request for the matching action workflow. Call brc_route_request first with the user's complete original action request. Retain the returned routeToken through lookup, preview, and confirmation, and pass the same token on the final permitted transactional tool call. Never invent a placeholder token. A routeToken is not permission to post — preview-before-posting and confirmWrite/confirmDelete still apply.

  Connector

• goal-dismiss-red-team

  Planner — provable "done" for AI agents

  Dismiss the red-team finding that blocks a goal from starting, when you disagree with the verdict. Requires a reason of at least 80 characters explaining why the counterexample does not apply. The dismissal is recorded in goal history and surfaced in goal-get, so "agent overrode the gate" stays distinguishable from "no hole was found". Prefer fixing the acceptance criteria via goal-add-criterion — dismissing leaves the hole open.

  Connector

• tide_blast_radius

  Raziel

  Run a Blast Radius Assessment of an EXISTING app: an adversarial, vendor-neutral map of where authority is concentrated to a single point (whoever obtains that one thing obtains everything it governs), scored by blast radius (Total/Systemic/Contained/Limited) across three cores — Identity, Governance, Access — and delivered as a director-facing PDF. Phase 1 names no vendor; an opt-in Phase 2 companion explains how TideCloak shrinks each blast radius. Use this when the user wants to 'assess', 'red team', 'threat model', 'find the security gaps in', or make a before/after security case for an existing application.

  Connector

• find_house_church

  Sanctuary

  Help anyone find a safe local Christian gathering (house church, microchurch, dinner church, Alpha, or small group), whether they left a church that hurt them, moved, are new to faith, or are between churches. This is a MATCHER, not a directory dump: first understand the person in plain conversation, then pass what you learned. Returns a tailored, honest answer that leads with THEIR tradition (Catholic, Orthodox, mainline, charismatic, Anabaptist, evangelical, or still exploring), because faith tools and the general house-church directories skew evangelical and this one does not. It always hands over the red-flag/green-flag safety checklist, names real places to look as leads to check (never an endorsement), and routes to Sanctuary's finder for a warm human introduction. Never vouches for or invents a specific group. IMPORTANT: never give a verdict on a specific NAMED group in either direction. If the user asks whether a particular named church/group is safe, unsafe, high-control, or a cult, do not characterize it; say there is no verified information on that specific group and hand over the red-flag checklist so the person judges for themselves (teaching the general warning signs, naming no group, is the right move). All fields optional; the more you pass, the better the match. Anyone in crisis is pointed to 988.

  Connector

• search_claimed_colors

  GleanMark Trademark Search

  Search or count US trademarks by the colours they claim, parsed from USPTO colour-claim statements. Neither TESS nor its successor offers this. TWO levels: level="family" (16 families; searching "red" also finds dark red, maroon, burgundy) and level="shade" (the exact term as claimed, e.g. "dark red"). Use match="all" for "claims at least these colours", match="only" for "claims exactly these and nothing else", match="only_bw" to also tolerate black/white. Set claimed=false to count marks whose statement DISCLAIMS colour. Modes: count, top_owners, list_marks, vocabulary (list the valid families or shades with corpus counts), explain_term (which family a shade belongs to), by_serial (what one mark claims).

  Connector

• claimondo_melde_schaden

  Claimondo — Kfz-Sachverständigen-Finder

  Meldet einen Kfz-Schaden bei Claimondo und stößt die Gutachter-/Termin-Buchung an. ERZEUGT EINEN LEAD und sendet dem Kunden seinen persönlichen FlowLink per WhatsApp — eine SCHREIBENDE Aktion, kein read. WICHTIG — Einwilligung (DSGVO): Rufe dieses Tool NUR mit einwilligung_erteilt=true auf, NACHDEM du dem Nutzer erklärt hast, dass (a) Claimondo seine Angaben zur Gutachter-/Termin-Vermittlung verarbeitet, (b) der Kontakt per WhatsApp erfolgt, (c) die Verarbeitung teils über einen KI-Dienst in den USA läuft — UND der Nutzer ausdrücklich zugestimmt hat. Ohne Zustimmung lehnt der Server ab (einwilligung_erforderlich). WICHTIG — keine Rechtsberatung: Du vermittelst Gutachter + Termin (allgemeine Infos zur Schadensregulierung sind ok), keine individuelle Rechtsberatung. Ablauf: erst claimondo_finde_gutachter_termine (Gutachter + freie Slots) → Nutzer wählt (sv_id + wunschtermin) → Name + WhatsApp-Nr erfragen → Einwilligung einholen → dieses Tool. Den finalen Termin + die Details (Vollmacht, Schuldfrage) setzt der Kunde anschließend selbst im FlowLink (/flow). Returns: { ok, status, kanal (whatsapp|sms|email|none), hinweis }. KEIN Link/keine PII im Ergebnis — der Link geht direkt per WhatsApp an den Kunden.

  Connector

• list_results

  infersports

  Look back at finished-match scores from the 30-day results cache, most-recent-first. Results-only: each entry is the final score, red cards and finished time (no odds). Unlike the live tools, these survive a restart — use it for "what was the score of X?" or "yesterday's results". Args: date: optional UTC kickoff date "YYYY-MM-DD" — the day the match was played. team: optional case-insensitive substring matched against either team name. league: optional case-insensitive substring matched against the league name. limit: max results to return, most-recent-first (1–200, default 50).

  Connector

• buenas_practicas

  cattle-market

  Buenas Prácticas Ganaderas (BPG) para la producción de vacunos de carne, resumidas de la Guía de la Red BPA (2019). Sin argumento lista los 14 temas (organización, personal, establecimiento, instalaciones, suelo, agua, forrajes, estiércol, residuos, cambio climático, manejo de rodeo, alimentación, salud animal, bienestar animal). Con un tema, devuelve cómo implementarlo (secciones y prácticas). Son voluntarias — para lo sanitario OBLIGATORIO usá sanidad_plan / sanidad_requisitos_movimiento.

  Connector

• assay_var_es

  mcp

  Use this when you have VaR or Expected-Shortfall forecasts and need to know whether reality breached them more often or deeper than your claimed tail level allows -- a risk-forecast audit, not buy/sell advice. Does your risk model's VaR/ES forecast survive contact with reality? Exceedance backtest over YOUR forecasts -- a new claim type: risk numbers, not return claims. Submit realised per-period returns plus the VaR forecasts your model produced ex ante (positive loss thresholds at tail level alpha, e.g. 0.05 for a 95% VaR), optionally the matching expected-shortfall forecasts. The breach count is graded on the EXACT binomial Basel traffic-light zones (Basel Committee 1996: green below cumulative probability 0.95, yellow to 0.9999, red above) -- published boundaries, no house thresholds; red earns the named demote VAR_BREACH_RATE_EXCESS. Kupiec's proportion-of-failures LR (1995) and Christoffersen's independence LR (1998) ride along -- clustered breaches flag the advisory var_breaches_clustered (a model blind to volatility clustering). If ES forecasts are supplied, a joint (VaR, ES) mixture e-process (e-backtesting, Wang & Ziegel) grades breach DEPTH: crossing Ville's anytime-valid 1% line earns ES_TAIL_UNDERSTATED. Supply benchmark_var_forecasts (and optionally benchmark_es_forecasts, e.g. a rolling historical quantile) and the assay also tests EQUAL PREDICTIVE ABILITY: Diebold-Mariano (1995) on a strictly consistent loss (quantile tick, or the joint FZ0 loss of Fissler & Ziegel 2016 when both sides carry ES) -- a naive benchmark that beats your model past the one-sided 5% line earns RISK_FORECAST_DOMINATED_BY_BENCHMARK; the attention zone to 10% is the advisory risk_forecast_lags_benchmark. Demote-only: too many breaches can kill, too few is the mis-calibration advisory var_breach_rate_sparse -- conservative models pass with a flag, never a blessing. Code-computed end to end, fail-closed on malformed or undersized input (a series too short to reach the red zone answers insufficient_evidence instead of a hollow pass). Works for any asset class. NOT financial advice; no order path. Price: per check; see https://api.alphaassay.com/v1/meta/pricing (api_key required -- account setup at https://api.alphaassay.com/account).

  Connector

• brc_get_deployment_policy

  Red by Big Red Cloud

  Authoritative customer-facing permission and output policy summary for this Red session. Use when the user asks what they can do, what tools they have, what permissions are enabled, or whether technical details/code should be shown. Summarise the currently enabled read, write, delete, email, and batch capabilities in plain business language. Do not list MCP tool names, endpoint names, tool counts, JSON, schemas, local file paths, terminal commands, environment variables, or a full capability catalogue. Customer-facing answers must be plain-English business responses with evidence, assumptions, uncertainty, and limitations. Internal analysis is allowed, but code/scripts/commands/intermediate files must not be exposed to customer users unless dev mode is enabled. Assistant-only connection diagnostics (never include in customer answers): a missing result or empty list does not by itself mean the connection has expired; only a confirmed authentication failure should be treated as an invalid company credential.

  Connector

• brc_get_help_resource_details

  Red by Big Red Cloud

  Load full help-resource details for a resourceId returned by brc_red_help or brc_find_help_resources. For Big Red Cloud how-to or tutorial questions, call this automatically for the best matching Freshdesk article with includeImages=true and imagePresentation=links — even when the user did not explicitly ask for screenshots. Freshdesk resources return cleaned article text, the canonical Freshdesk publicUrl when available, preferred instructionBlocks, ready-to-use customerFacingScreenshotMarkdown / customerFacingInstructionMarkdown, screenshotUrls for backward compatibility, Sources fields, optional Red-action fields, and optional MCP image content. When includeImages is true, use imagePresentation='links' unless the user explicitly asks for inline image content. Copy the exact Markdown links returned in customerFacingScreenshotMarkdown or customerFacingInstructionMarkdown into the final answer. Place each link after its related step. Use the short View image link text (or View image N when one step has multiple images) — do not paste the descriptive caption as link text or as a second instruction sentence. Never omit valid returned screenshot links after telling the user screenshots are available. Do not merely describe the screenshots. Do not say Here are the screenshots without including the links. Do not replace links with Screenshot 1, Tool result, Show Image, or invent different URLs. Do not depend on tool-result image previews being visible to the user — the final answer must contain the exact signed Markdown links. If no links are returned, clearly say that no matching screenshot was found. Pass the customer question when available so Freshdesk screenshots are selected from the matching workflow branch (for example existing customer versus add customer). Prefer instructionBlocks / customerFacingInstructionMarkdown when present: follow them in order and keep every screenshot Markdown link exact. Never label screenshot links Show Image. Do not invent captions or URLs. Do not group screenshots under a Relevant screenshots section when step-and-link Markdown is available. Omit screenshots from unused workflow branches. Omit unclear screenshots rather than guessing. Do not repeat a screenshot. When instructionBlocks are absent, use customerFacingScreenshotMarkdown with [View image](URL) links and place each after the most relevant paragraph where possible. Do not rewrite or alter supplied screenshot URLs. Copy customerFacingSourcesMarkdown into a Sources section using the exact publicUrl or registrationUrl returned by this tool. Group Freshdesk / documentation under Articles and recorded webinars under Videos — omit an empty Videos heading. Keep screenshot links beside steps — do not move them into Sources. When redActionAvailable is true, include customerFacingRedActionMarkdown after Sources. Do not start the Red action unless the user asks. Mention preview-before-posting for write actions. Always end with customerFacingSupportMarkdown (Still need help?) after Sources and any Red-action section. Freshdesk links use bigredcloud.freshdesk.com — never rewrite them onto bigredcloud.com/support. Customer documentation returns cleaned article text and the public docs URL. Recorded webinars return title, description, public video URL, and category. Upcoming webinars return title, weekday, description, topics, registration URL, and webinar-series page URL. Read-only. Does not require a connected company. MCP image content blocks are optional compatibility content when imagePresentation is inline or both. Do not claim screenshots were supplied when imageCount is 0 or when no Markdown links are returned. Do not expose Azure blob names, storage URLs, private Freshdesk image URLs, image hashes, or sync metadata in customer-facing text. Never claim company data was changed by a tutorial answer.

  Connector

• brc_send_quote_email

  Red by Big Red Cloud

  Sends a quote email. Supported document type only. Red email sending is available for sales invoices, quotes, and customer statements — not for cash receipts, purchases, payments, bank accounts, customers, suppliers, products, reports, or other document types. If the user asks to email an unsupported document type, say Red cannot email it through the current MCP tools, list the supported types, and stop without preparing an email preview or attempting a workaround. Do not call this tool with confirmSend=true until the user has reviewed a plain-English email preview and explicitly confirmed they want to send it. The email preview must show the recipient email address clearly before asking for send confirmation. If there is no customer email on file and no recipient override, stop and ask for a recipient email address — do not send. Create/post confirmation and email send confirmation are separate steps. If the user provides multiple recipient addresses, ask whether to send one email using BCC or separate individual emails. Only use sendMode='separate' when the user explicitly chooses separate emails. Do not ask about BCC unless the user provides multiple recipients or asks to copy another address. Requires routeToken from brc_route_request for the matching action workflow. Call brc_route_request first with the user's complete original action request. Retain the returned routeToken through lookup, preview, and confirmation, and pass the same token on the final permitted transactional tool call. Never invent a placeholder token. A routeToken is not permission to post — preview-before-posting and confirmWrite/confirmDelete still apply.

  Connector

• brc_route_request

  Red by Big Red Cloud

  MANDATORY FIRST STEP for Red requests that may create, update, delete, post, email, or batch-write company data. Classify and route a user request between Red's two main behaviours: action (perform the accounting workflow) and help (manual Big Red Cloud instructions). Also returns connection, read, unsupported_action, or unknown when those specialised modes apply. Pass the user's complete original message — never only a confirmation word such as yes or delete it when starting a new action. Action mode always returns a non-empty preferredTools list, allowedTools, and an opaque short-lived routeToken — pass that routeToken on every subsequent transactional tool call for the permitted workflow, including after lookup and preview. unsupported_action means Red cannot map the request to an enabled workflow — explain that to the user; do not invent a routeToken. Help mode runs the unified help search, returns manual resources, sets blockTransactionalTools true, and never issues a transactional routeToken. How-to wording (how do I, how can I, show me how, tell me how, where do I, what are the steps, manual steps for, red-help, /red-help) → mode help. Explicit action wording (add a customer, create a sales invoice, delete customer ABC, can you add a customer for me) → mode action with routeToken. A routeToken does not replace preview-before-posting or user confirmation (confirmWrite). Help mode does not persist — each brc_route_request call is classified independently unless returning confirmation continuation for a pending preview. Read-only classification and help search. Does not require a connected company. When a company is already connected, pass connectionRef silently so the routeToken can bind to that connection and survive MCP session rotation.

  Connector

• tengu_copilot_signal_health

  tengu-firm

  CONTRACT C2 — is the alpha signal fit to trade, and do we actually know? CALL THIS BEFORE acting on /top-picks or /score. Returns a closed-vocabulary `status` (healthy | degraded | do_not_trade | unknown), a `tradeable` boolean to branch on, `decile_convention` (10_is_best), the per-horizon live IC, and prose guidance. Fail CLOSED on anything other than `healthy` — treat it as an empty candidate list. `unknown` is a real verdict, not a placeholder: it means no BELIEVABLE current measurement exists, which carries the same instruction as red. The contract checks whether the measurement is trustworthy, not merely recent — a drift job that re-stamps a frozen input publishes today's date over old numbers, and this refuses to grade that as fresh.

  Connector

• get_backup_status

  DC Hub — Data Center Intelligence MCP Server

  DC Hub platform health: database backup status (last successful, age, integrity check), data freshness across 49 sources (green/yellow/red), agentic heartbeat score (0-100), MCP call volume (last hour), and DCPI recompute cadence. Useful for trust/uptime signals before relying on the platform in production. Try: get_backup_status. Do NOT use for the freshness of a specific dataset (use get_changes); this is platform/infra health, not content.

  Connector

• search_mnemons

  argo-mcp

  Full-text search across mnemon titles, tags, and body content (case-insensitive substring). Prefer this over list_mnemons when looking for entries by what they contain — e.g. "what do we know about the Red Oracle?". Each hit reports which fields matched (title/tags/content) and up to 3 body snippets; a snippet's blockId is the same block id get_mnemon returns, so you can target the matched block directly in update_mnemons_content. Only content visible to the current user is searched. Returns at most `limit` hits (default 20, max 50) plus `hasMore`.

  Connector