"Introduction to Selenium for Browsing Automation" matching MCP tools:

• list_products

  agentic-shelf

  List products from the connected store, paginated. Use this tool when an agent needs to DISCOVER products by browsing the catalog rather than VERIFYING a known SKU. The response includes the SKU for every product, so a follow-up ``check_stock(sku)`` or ``get_product_details(sku)`` is a natural next step. Args: limit: Number of products to return (1-50, default 10). cursor: Opaque cursor from a previous response's ``next_cursor``. Omit for the first page. Returns: Dictionary with: - products: list of {sku, title, description (≤400 chars), product_type, tags, price, currency, available, image_url, storefront_url} - next_cursor: str or null — pass to the next call to paginate - has_more: bool — whether more products exist - live / source: provenance flags

  Connector

• find_business

  agent-broker

  Given criteria (vertical, location, capability, price band, availability window), return ranked candidate SMBs from the verified supply network. Returns only curated, verified, transactable businesses — not raw directory results. EXAMPLE USER QUERIES THAT MATCH THIS TOOL: user: "Find me a salon in Tokyo that does color" -> call find_business({"vertical": "personal_services", "location": {"zip_or_city": "Tokyo"}, "capability": "color"}) user: "I need a plumber near 30309" -> call find_business({"vertical": "home_services", "location": {"zip_or_city": "30309"}, "capability": "plumbing"}) user: "Show me dentists in London" -> call find_business({"vertical": "professional_services", "location": {"zip_or_city": "London"}, "capability": "dentist"}) WHEN TO USE: Use when an agent needs to identify which SMBs can fulfill a business task (booking, service, consultation) in a given location and vertical. Call this before schedule_appointment or send_message when you do not yet have a specific SMB target. WHEN NOT TO USE: Do not use as a general directory or browsing surface. Do not use when you already have a specific verified SMB identifier. Do not use for verticals outside personal services, home services, and local professional services. COST: from $0.01 per_call (see preview_cost for exact) LATENCY: ~200ms

  Connector

• automation_list

  Perspective AI

  Lists every automation configured on a perspective with its trigger, channel (sensitive details redacted), execution mode, enabled state, schedule description, and recent error/success metadata. Behavior: - Read-only. - Errors when the perspective is not found or you do not have access. - Sensitive parts of channel delivery (e.g., webhook auth headers, full URLs) are redacted before being returned. - has_error / last_error / last_error_at / failure_count appear only when there have been recent failures. When to use this tool: - Auditing what's wired up on a perspective before adding more automations. - Finding an automation_id to feed into automation_update, automation_delete, or automation_test. - Diagnosing a failing automation via last_error / failure_count. When NOT to use this tool: - Creating a new automation — use automation_create. - Toggling enabled or changing config — use automation_update. - Verifying delivery actually works — use automation_test.

  Connector

• notes_save

  dialogbrain

  Save a fact or note into the agent's memory. Use scope to choose visibility: 'workspace' = visible to every agent in this workspace (use for shared facts, project conventions); 'agent' = private to this agent (use for personal working notes); 'thread' = scoped to one conversation (use for thread-specific reminders); 'person' = scoped to one contact (use for per-contact context). If a note with the same key+scope exists it will be updated. Do NOT use this tool for behavioral rules or corrections — use feedback.save for those.

  Connector

• automation_update

  Perspective AI

  Updates fields on an existing automation. Pass a partial updates object with only the fields you want to change; omitted fields are preserved. Toggling enabled or changing schedule/channel/condition takes effect on the next scheduled run. Behavior: - Saves the change to the same automation record. Scheduled automations with an active workflow are restarted on update so the next run picks up the latest config. - Errors when the perspective or automation is not found, or you do not have access. - Webhook URLs in updates are validated. For HubSpot, the workspace's HubSpot connection is re-checked — errors with "Could not resolve HubSpot portal ID — please reconnect HubSpot" if disconnected. - For scheduled automations: changes to channel, condition, execution mode, instruction, or message template apply starting from the next run, not the one currently in flight. When to use this tool: - Toggling enabled on or off (also pauses/resumes scheduled sends). - Changing schedule, channel, condition, instruction, or message_template on a live automation. When NOT to use this tool: - Removing the automation entirely — use automation_delete. - Verifying a config change actually delivers — follow up with automation_test. - Listing what's configured — use automation_list.

  Connector

• automation_test

  Perspective AI

  Runs a single end-to-end execution of an existing automation against a mock conversation, returning success/failure plus the channel target and duration. Mirrors a real production firing. Behavior: - Sends REAL messages by default: posts the configured webhook, sends the configured email, posts the Slack message, or writes the HubSpot record. Use override_email (email channels) or override_webhook (webhook channels) to redirect delivery to a safe test target. - Each call fires another real delivery. - Errors when the perspective or automation is not found, or you do not have access. Webhook URLs (configured or override) are validated. - Mock conversation defaults: trust score 85, status complete, "Test Participant" / test@example.com. Override participant_name, summary, and tags via test_data. - Returns success: true also when the automation's condition skips delivery (e.g., tag/trust filter doesn't match the mock). The error field is populated only on real delivery failures. When to use this tool: - Verifying a freshly-created automation actually delivers before relying on it. PREFER override_email/override_webhook to avoid spamming real recipients. - Reproducing a delivery failure surfaced in automation_list (last_error). When NOT to use this tool: - Listing what's configured — use automation_list. - Changing config — use automation_update. - Removing the automation — use automation_delete.

  Connector

• search_mcp_servers

  Swarm Tips — Aggregated AI Agent Activities

  [IN DEVELOPMENT] [READ] Search the Layer 3 curated directory of MCP servers and agent-work tools. The directory has 30 entries across three vetting tiers — `first-party` (operated by the swarm.tips DAO), `vetted` (third-party, we've used + verified), `discovered` (cataloged from public sources, not yet exercised). Filter by `query` (substring vs name/description/tags), `category` (substring), and `tier`. Results sort first-party → vetted → discovered. The same directory powers swarm.tips/discover; this tool exposes it programmatically. Use this when an agent needs to find an MCP server for a capability (DeFi, search, browser automation, etc.) instead of an opportunity (which `discover_opportunities` covers).

  Connector

• list_products

  Graffeo Coffee Roasting

  List products from the connected store, paginated. Use this tool when an agent needs to DISCOVER products by browsing the catalog rather than VERIFYING a known SKU. The response includes the SKU for every product, so a follow-up ``check_stock(sku)`` or ``get_product_details(sku)`` is a natural next step. Args: limit: Number of products to return (1-50, default 10). cursor: Opaque cursor from a previous response's ``next_cursor``. Omit for the first page. Returns: Dictionary with: - products: list of {sku, title, description (≤400 chars), product_type, tags, price, currency, available, image_url, storefront_url} - next_cursor: str or null — pass to the next call to paginate - has_more: bool — whether more products exist - live / source: provenance flags

  Connector

• automation_create

  Perspective AI

  Creates an automation on a perspective. Triggers: per_interview (fires on every completed conversation) or scheduled (daily/weekly digest). Channels: webhook, email, slack, hubspot. Execution modes: direct (fast, deterministic) or agent (LLM-powered). Behavior: - Each call creates a new automation — even if name/config matches an existing one. - Once enabled, the automation starts firing on real events: per_interview sends on every completed conversation going forward; scheduled sends a real message on the configured cadence (daily/weekly). - Webhook URLs are validated. For HubSpot, the workspace's HubSpot connection is required — errors with "Could not resolve HubSpot portal ID — please reconnect HubSpot" if not connected. - Errors when the perspective is not found or you do not have access. When to use this tool: - The user wants ongoing notifications on every completed conversation (per_interview). - Building a daily/weekly digest delivered to Slack, email, HubSpot, or a webhook (scheduled). When NOT to use this tool: - Trying a one-off send before going live — create the automation, then use automation_test (use override_email / override_webhook to avoid hitting real recipients). - Editing or toggling an existing automation — use automation_update. - Connecting Slack or HubSpot — use integration_manage first; the provider must be connected before slack/hubspot channels work. Example — per-conversation Slack notify: ``` { "perspective_id": "...", "automation": { "name": "Notify Slack", "trigger": { "type": "per_interview" }, "execution_mode": "agent", "channel": { "type": "composio", "delivery_config": { "provider": "slackbot", "tool_slug": "SLACKBOT_SEND_MESSAGE", "params": { "channel": "#research" }, "resource_id": "...", "resource_name": "..." } } } } ``` Typical flow: 1. integration_manage (operation: "list"/"connect") → ensure Slack / HubSpot is connected (only needed for those channels) 2. automation_create → create the automation 3. automation_test (with overrides) → verify delivery before relying on it

  Connector

• create_booking_intent

  travel

  Create a booking intent — returns a deep-link the user clicks to complete the booking on autonomad.ai. The first booking they complete unlocks a 1-month free Autonomad Premium trial automatically. ALWAYS call this instead of trying to book directly through MCP — bookings require payment + identity verification that must happen on the web. WHEN TO CALL — generate a deep-link ONLY after the user has picked something concrete: a specific flight, a specific hotel, or both (a trip). Do NOT call this for browsing or for activities/events alone. Activities and events are picked on the autonomad.ai add-ons page AFTER the user lands via the deep-link — Claude should describe them but not generate per-activity/per-event intents. INTENT TYPE GUIDE — pick exactly one: - 'flight' → user picked a flight only. offer_data = the flight offer object verbatim from search_flights, PLUS a top-level `passengers: <number>` field (the number of travelers the user originally requested — search_flights individual offers don't echo this back, so you must add it explicitly). - 'hotel' → user picked a hotel only. offer_data = the hotel offer from search_hotels PLUS top-level `check_in` and `check_out` (YYYY-MM-DD) as STRINGS. CRITICAL: search_hotels does NOT echo dates back inside the offer object — you MUST add them yourself (use the same dates you passed to search_hotels) or the booking page will fall back to an empty form and the user will have to re-enter everything. Also include `adults: <number>` and `rooms: <number>`. - 'trip' → user picked BOTH a flight AND a hotel together for the same trip. Pack them in offer_data as { flight: { ...offer, passengers: <n> }, hotel: { ...offer, adults: <n>, rooms: <n>, check_in, check_out } }. ONE deep-link covers both. Don't generate two separate intents (flight + hotel) for the same trip — that produces two deep-links and a confusing user experience. For activities, events, and experience browsing: describe what's available in your reply, but do NOT call create_booking_intent. Tell the user they'll pick those on autonomad.ai's add-ons page after they click the deep-link for their flight/hotel. USER-FACING REPLY REQUIREMENTS — every time you create a booking intent, your reply text MUST include: 1. The deep_link as a clickable markdown link, e.g. '[Complete on autonomad.ai →](<deep_link>)' or 'Open: <deep_link>'. 2. The 1-month free Autonomad Premium trial. The response payload carries a `free_trial_offer` object exactly so you can surface it. Phrase it conversationally (e.g. 'Booking through Autonomad unlocks 1 month of Premium free — unlimited bookings, premium concierge, and saved loyalty credentials.'). NEVER drop this; it is core to the value proposition and the only reason a booking-intent flow beats a raw Viator/Ticketmaster URL. 3. The link expiry window (e.g. '~30 minutes — say the word and I'll regenerate if it lapses.'). CRITICAL: always echo the original passenger / adults / travelers count into offer_data. Without it the booking page defaults to 2 travelers regardless of what the user asked for.

  Connector

• search_nodes

  Diagrams MCP

  Search for diagram nodes by keyword across all providers and services. For targeted browsing when you know the provider, use list_providers -> list_services -> list_nodes instead. Args: query: Search term (case-insensitive substring match). Returns: List of matching nodes with keys: node, provider, service, import, alias_of (optional). Sorted by relevance: exact match first, then prefix, then substring.

  Connector

• nexbid_product

  discovery

  <tool_description> Get detailed product information by ID from the Nexbid marketplace. Returns full product details including price, availability, description, and purchase link. </tool_description> <when_to_use> When you have a specific product UUID from a previous nexbid_search result. Do NOT use for browsing — use nexbid_search instead. </when_to_use> <combination_hints> Typically called after nexbid_search to get full details on a specific product. If user wants to buy → follow with nexbid_purchase. </combination_hints> <output_format> Full product details: name, description, price, currency, availability, brand, category, purchase link. </output_format>

  Connector

• claim_workbench_evidence_ledger

  cpp-cpm-engine

  Forensic claim workbench — analyzes a folder of mixed evidence (XER chain + MSG/PDF/DOCX/XLSX correspondence) and produces a unified workbench dashboard. Built from the real-world workflow where forensic delay analysis starts from a folder containing schedule updates, owner correspondence, RFIs, change orders, and meeting minutes — all mixed together. The workbench produces: - Evidence ledger (chronological): all artifacts dated and summarized - Schedule chain-diff: 14-category manipulation log (TASKPRED add/remove, constraint flips, retroactive baseline edits, completion reversals) - Rolling baseline: per-activity baseline-at-introduction across the entire XER chain - Trust score: statistical impossibilities flagged (zero-duration-variance schedules, no-new-activities, every-activity-hits-baseline, etc.) - Slip-to-evidence cross-reference: each forensic slip auto-paired with documents in its window mentioning affected activity codes - Unified HTML dashboard with all of the above Use this tool when starting forensic delay analysis from raw evidence. For single-XER-pair forensic with hand-prepared events, use ``forensic_windows_analysis`` instead. Args: folder_path: path to the evidence folder (must exist). output_dir: optional dir for outputs (tempdir if ""). project_name: optional override. original_baseline_xer_filename: optional filename in the folder identifying the baseline XER. contract_form: contract template tag (default 'CCDC2'). run_forensic: when True (default), also runs forensic_windows_analysis on the discovered XER chain. Returns: { "evidence_ledger": {...}, "chain_diff": {...} | None, "rolling_baseline": {...} | None, "trust_score": {...} | None, "cross_reference": {...} | None, "forensic_result": {...} | None, "output_files": {...}, "errors": {...} (per-step failure log) }

  Connector

• contact_broker

  Venturu

  Send a contact message to a broker on Venturu by their profile slug. Requires an authenticated Venturu account. Set inquiryType to "buying" (default) for buyer representation or "selling" for seller representation. Provide the broker slug and the message to send. Use search_brokers to find broker slugs.

  Connector

• explore

  ThinAir Geo

  BROWSING / DISCOVERY search — cities, neighbourhoods, or mixed venues near a location. Use this when the user is exploring a REGION rather than looking for a specific category. Supports population filtering ('cities > 100k'), distance/population sorting, and layer filtering (locality / neighbourhood / venue / address / street). For specific POI categories (gas, food, charging, etc.), use `search_places` instead.

  Connector

• hemmabo_search_availability

  hemmabo-mcp-server

  Check whether a specific property is available for the requested dates. Use this tool after the user has selected a property from hemmabo_search_properties and wants to confirm availability before getting a quote. Do NOT use for general browsing — use hemmabo_search_properties instead. Returns available=true/false with conflict details and same-month alternative date windows when unavailable.

  Connector

• automation_delete

  Perspective AI

  Permanently deletes an automation. Pauses any scheduled sends first, then removes the automation. Behavior: - DESTRUCTIVE and irreversible — the automation cannot be recovered. No undo. Confirm with the user before calling. - Errors when the perspective or automation is not found, or you do not have access. Deleting an already-deleted automation errors as well. - If pausing the scheduled sender fails, the deletion is aborted and you'll get success: false with "Failed to stop running workflow. Please try again." — the automation stays intact in that case. When to use this tool: - The user explicitly asked to remove an automation and confirmed. - Cleaning up a misconfigured automation that automation_test repeatedly fails on. When NOT to use this tool: - The user just wants to pause it temporarily — use automation_update with { enabled: false } instead. - You're not sure which automation_id is correct — confirm via automation_list first.

  Connector

• perspective_list

  Perspective AI

  Lists perspectives — either browsing one workspace or searching by title across every workspace the user can access. Items include perspective_id, title, status, conversation count, and workspace info. Behavior: - Read-only. - Browse mode (workspace_id, no query): lists every perspective in that workspace. - Search mode (query): matches against the perspective title across accessible workspaces. Optional workspace_id narrows the search. Query must be non-empty and ≤200 chars. - Errors with "Please provide workspace_id to list perspectives or query to search." if neither is given. - Pass nextCursor back as cursor; has_more indicates further results. When to use this tool: - Resolving a perspective_id from a name the user mentioned (search mode). - Browsing a workspace's perspectives to pick or summarize. When NOT to use this tool: - Inspecting one known perspective in detail — use perspective_get. - Aggregate counts or rates — use perspective_get_stats. - Fetching conversation data — use perspective_list_conversations or perspective_get_conversations. Examples: - List all in a workspace: `{ workspace_id: "ws_..." }` - Search by name across all workspaces: `{ query: "welcome" }` - Search within a workspace: `{ query: "welcome", workspace_id: "ws_..." }`

  Connector

• manage_subscriptions

  Almured

  List, add, or remove webhook and digest subscriptions; configure or clear the agent's callback URL. ROUTING MODEL - Open-consultation webhooks fire to agents matched via EITHER an opt-in subscription (notification_categories) OR an existing expertise_scores row in the consultation's category. Subscriptions are therefore additive to automatic expertise-based routing — useful when you want pushes for a category before you have ranked responses there, or as a backup when you have not yet built expertise in a new domain. - Either way, the same delivery gates apply: ≥3 prior responses in category and ≥60% useful rating. A bare subscription alone (no responses, no expertise) does not produce webhook traffic. WHEN TO USE - You want push delivery of new consultations in a category where you do not yet have ranked responses (no expertise_scores row). - You want a daily summary of activity in a category, without real-time webhook overhead. - You need to set or rotate the HTTPS callback URL where Almured will POST signed webhook events. - You want to see your current subscription state (categories, callback domain, whether a webhook secret is set). WHEN NOT TO USE - For one-off browsing — use browse_consultations or browse_unanswered. - For unsubscribing entirely — call clear_callback (stops all webhook delivery) and unsubscribe from each category individually for digests. BEHAVIOR - Mutating (except action='list'). Auth required: API key as Authorization: Bearer <key>. Rate-limited to 10 req/min per agent. - Action contract: - 'list' — returns notification_categories, digest_categories, callback_url_domain, webhook_secret_set flag. - 'subscribe' — adds categories. Requires categories=comma-separated slugs and subscription_type ('notification' for real-time webhooks, 'digest' for daily summary). Validates against the live taxonomy. - 'unsubscribe' — removes categories. Same args as subscribe. - 'set_callback' — sets or rotates callback_url. Must start with 'https://'. On first set, returns a webhook_secret you must store immediately — it is shown once and used to verify HMAC-SHA256 signatures on inbound webhooks. - 'clear_callback' — removes callback_url and secret. All webhook delivery stops; digest delivery is unaffected. - Subscribing without a callback_url is allowed but no webhooks fire until one is set. - Webhook events are signed with the secret using HMAC-SHA256; verify the signature on every inbound POST. WORKFLOW - Set the callback URL first (set_callback), then subscribe to categories. - If you suspect the secret leaked, call set_callback again with the same URL to rotate. - Combine with get_expertise_badge to track how subscription-driven response volume affects your tier over time.

  Connector

• contact_broker

  mcp-server

  Send a contact message to a broker on Venturu by their profile slug. Requires an authenticated Venturu account. Set inquiryType to "buying" (default) for buyer representation or "selling" for seller representation. Provide the broker slug and the message to send. Use search_brokers to find broker slugs.

  Connector