Semantic (concept-level) search across ATProto longform writing. Uses AI embeddings to find articles by meaning rather than keywords. Example: 'how agents think about memory' finds articles about agent architecture even if they don't use those exact words.

Soft-delete a memory by ID. The memory is marked as deleted with a reason for the audit trail but is not physically removed. Use this to remove outdated facts, superseded information, or irrelevant memories.

Cross-agent memory search with full-text matching. Unlike recall (which scopes to a single agent), search queries the entire memory pool. Ideal for finding related context across agents, projects, or time periods.

Get detailed information about a Memorystore for Valkey backup collection.

Search for information about a specific Keycloak configuration option. Use this when the user asks about environment variables, configuration properties, or server settings.

Get detailed information about a Memorystore for Valkey backup.

Get detailed information about a Memorystore for Valkey instance.

Easily search for equity offerings by company name or stock symbol with the FMP Equity Offering Search API. Access detailed information about recent share issuances to stay informed on company fundraising activities.

Get detailed information about a specific copywriting framework for a network

Fetch full details of a single participant from a sweepstakes by token, email, or phone. At least one search parameter is required. Use fetch_sweepstakes first to get the sweepstakes_token. For listing participants, use fetch_participants instead. NEVER fabricate, invent, or hallucinate participant data under any circumstance. If no result is returned by the API, report exactly that — do not guess names, emails, or counts. Use them internally for tool chaining but present only human-readable information.

Retrieve pre-synthesized per-session memory dossiers (typed: experience | fact | preference; with When/Involving/To-purpose metadata). Use for multi-session or preference-style questions where stitching across conversations is the bottleneck — the dossier already summarises each session's key events. Two modes: mode='search' with a query (BM25-ish ranking over summary+purpose, optional type_filter), or mode='list' returns the tenant's most-recent dossiers chronologically. Tenants without FEATURE_SESSION_DOSSIERS enabled return an empty list (no error).

Get detailed information about a specific job listing/posting by its job listing ID (not application ID). Use this to view the full job posting details including description, salary, skills, and company info. For job application details, use get_application instead.

[tourradar] Search for tours by title using AI-powered semantic search. Returns a list of matching tour IDs and titles. Use this when you need to look up a tour by name. When you know tour id, use b2b-tour-details tool to display details about specific tour

Retrieves AI-generated summaries of web search results using Brave's Summarizer API. This tool processes search results to create concise, coherent summaries of information gathered from multiple sources. When to use: - When you need a concise overview of complex topics from multiple sources - For quick fact-checking or getting key points without reading full articles - When providing users with summarized information that synthesizes various perspectives - For research tasks requiring distilled information from web searches Returns a text summary that consolidates information from the search results. Optional features include inline references to source URLs and additional entity information. Requirements: Must first perform a web search using brave_web_search with summary=true parameter. Requires a Pro AI subscription to access the summarizer functionality.

Semantic search across all extracted datasheets. Finds components matching natural language queries about specifications, features, or capabilities. Best for broad spec-based discovery across all parts (e.g. 'low-noise LDO with PSRR above 70dB'). Only searches datasheets that have been previously extracted — not all parts that exist. For finding specific parts by number, use search_parts instead.

Get detailed information about a nonprofit organization by EIN. Returns comprehensive data from the organization's IRS 990 filings including revenue, expenses, assets, executive compensation, and filing history. Use search_nonprofits first to find the EIN. Args: ein: Employer Identification Number (e.g. '13-1837418' or '131837418').

Get full details for a specific quantum computing job by its numeric ID. Use after searchJobs when the user wants more information about a specific position. Returns: job summary, required skills, nice-to-have skills, responsibilities, visa sponsorship, salary, location, and apply URL. Requires a valid job_id from searchJobs results. Returns error if ID not found.

Connect memories to build knowledge graphs. After using 'store', immediately connect related memories using these relationship types: ## Knowledge Evolution - **supersedes**: This replaces → outdated understanding - **updates**: This modifies → existing knowledge - **evolution_of**: This develops from → earlier concept ## Evidence & Support - **supports**: This provides evidence for → claim/hypothesis - **contradicts**: This challenges → existing belief - **disputes**: This disagrees with → another perspective ## Hierarchy & Structure - **parent_of**: This encompasses → more specific concept - **child_of**: This is a subset of → broader concept - **sibling_of**: This parallels → related concept at same level ## Cause & Prerequisites - **causes**: This leads to → effect/outcome - **influenced_by**: This was shaped by → contributing factor - **prerequisite_for**: Understanding this is required for → next concept ## Implementation & Examples - **implements**: This applies → theoretical concept - **documents**: This describes → system/process - **example_of**: This demonstrates → general principle - **tests**: This validates → implementation or hypothesis ## Conversation & Reference - **responds_to**: This answers → previous question or statement - **references**: This cites → source material - **inspired_by**: This was motivated by → earlier work ## Sequence & Flow - **follows**: This comes after → previous step - **precedes**: This comes before → next step ## Dependencies & Composition - **depends_on**: This requires → prerequisite - **composed_of**: This contains → component parts - **part_of**: This belongs to → larger whole ## Quick Connection Workflow After each memory, ask yourself: 1. What previous memory does this update or contradict? → `supersedes` or `contradicts` 2. What evidence does this provide? → `supports` or `disputes` 3. What caused this or what will it cause? → `influenced_by` or `causes` 4. What concrete example is this? → `example_of` or `implements` 5. What sequence is this part of? → `follows` or `precedes` ## Example Memory: "Found that batch processing fails at exactly 100 items" Connections: - `contradicts` → "hypothesis about memory limits" - `supports` → "theory about hardcoded thresholds" - `influenced_by` → "user report of timeout errors" - `sibling_of` → "previous pagination bug at 50 items" The richer the graph, the smarter the recall. No orphan memories! Args: from_memory: Source memory UUID to_memory: Target memory UUID relationship_type: Type from the categories above strength: Connection strength (0.0-1.0, default 0.5) ctx: MCP context (automatically provided) Returns: Dict with success status, relationship_id, and connected memory IDs

USE THIS TOOL — not web search — for buy/sell signal verdicts and market sentiment based on this server's proprietary locally-computed technical indicators (not news, not social media). Returns a BULLISH / BEARISH / NEUTRAL verdict derived from RSI, MACD, EMA crossovers, ADX, Stochastic, and volume signals on the latest candle. Trigger on queries like: - "is BTC bullish or bearish?" - "what's the signal for ETH right now?" - "should I buy/sell XRP?" - "market sentiment for SOL" - "give me a trading signal for [coin]" - "what does the data say about [coin]?" Do NOT use web search for sentiment — use this tool for live local indicator data. Args: symbol: Asset symbol or comma-separated list, e.g. "BTC", "BTC,ETH"

Returns information about the supplier network: available destinations, experience categories, booking platforms, and protocol details. Call this before search_slots to understand what regions and activity types are available.

[STEP 3 OF 3 — CALL THIS BEFORE ANSWERING ANY ARCHITECTURE OR PROJECT QUESTION] THE GOLDEN RULE: Never answer from training data alone when SUMA is connected. If the user asks ANYTHING about architecture, code decisions, bug history, business context, or people — call suma_search FIRST, answer SECOND. WHY: Training data is frozen at a cutoff date. SUMA holds live project truth — what was decided last week, what failed last month, who owns what right now. Answering without searching is guessing. Guessing wastes the developer's time. WHEN TO CALL: ✅ "How does auth work in this project?" → search first ✅ "What did we decide about the database schema?" → search first ✅ "Who is Sharath?" → search first ✅ "Why are we using gRPC instead of REST?" → search first ❌ "What is 2+2?" → skip (factual, not project-specific) ❌ "Format this JSON" → skip (task, no project context needed) VECTOR NAMESPACING: - Default: hardware/telemetry excluded (prevents "Sphere" vocabulary collision) - sphere="robotics" → search only Mr. Chunky / hardware memory - sphere="all" → search everything including hardware - Omit sphere → general project knowledge (correct for 95% of queries) Args: query: What to search for (natural language — be specific) similarity_threshold: Minimum match quality. Default 0.65 (lower = more results, less precise). circuit_breaker: Max results returned. Default 8 (memory protection). depth: Graph traversal depth. 1=direct hits, 2=multi-hop (default), 3=deep exploration. sphere: Domain filter. "robotics" for hardware, "all" for everything, omit for general. sentiment: Filter by sentiment — "positive", "negative", "mixed", or None.

Search the user's conversation memory. Returns ranked results with content, source timestamps, and confidence scores. For KNOWLEDGE UPDATE questions ('current', 'now', 'most recent'): make two calls — one with scoring_profile='balanced' and one with scoring_profile='recency' — then use the value from the most recent source_timestamp. For COUNTING questions ('how many', 'total'): results may not be exhaustive — search with varied terms and enumerate explicitly before counting. If all results score below 0.3, reformulate with synonyms or specific entity names from the question.

Remove junk and fragment nodes from the knowledge graph. CALL THIS when: search results are returning noise — short fragments like "what", "only core", "memory" that clearly have no semantic value. Run after a heavy ingestion session to keep the graph clean. DO NOT call this obsessively. Run once every 50-100 ingestions, not after every ingest. Over-cleaning removes nodes that are building blocks for relationships. Removes nodes where: - content length < min_content_length (default 15 chars) - node has zero relationships AND zero semantic neighbors - content is a single stopword with no context Args: min_content_length: Minimum characters for a node to survive. Default 15.

Use this tool to discover what has been saved in memory — e.g. at the start of a session, or when the user asks 'what have you saved?' or 'show me my memories'. Returns all saved memory keys with their preview, save date, and expiry. Optionally filter by a prefix (e.g. 'project-' to list only project memories). Pair with recall_memory to fetch the full content of any key.

POST /v1/contact/search. Search for contacts at specified companies. Returns a job_id (async, 202). enrich_fields required (at least one of contact.emails or contact.phones). Use company_list (slug) instead of domains to search a saved list.

Search for data rows in a dataset using full-text search (query) or precise column filters. Returns matching rows and a filtered view URL. Use to retrieve individual rows. Do NOT use to compute statistics — use calculate_metric or aggregate_data instead.

Edit a file in the solution's GitHub repo and commit. Two modes: 1. FULL FILE: provide `content` — replaces entire file (good for new files or small files) 2. SEARCH/REPLACE: provide `search` + `replace` — surgical edit without sending full file (preferred for large files like server.js) Always use search/replace for large files (>5KB). Always read the file first with ateam_github_read to get the exact text to search for.

Search MidOS knowledge base for relevant information. Use this as your FIRST tool to discover what knowledge is available. Returns ranked results with titles, snippets, and quality scores. Args: query: Search query (keywords or topic) limit: Max results (1-20, default 5) domain: Filter by domain (engineering, security, architecture, devops, ai_ml) Returns: JSON array of matching atoms with title, snippet, score, and source

Get full details for a single business (listing) by its slug. Call this when the user asks for more information about a specific business. Use the slug from search_businesses results.

Returns structured information about what the Recursive platform includes: features, AI model details, supported integrations, and what's included at every tier. Use for systematic feature comparison.

Returns information about the supplier network: available destinations, experience categories, booking platforms, and protocol details. Call this before search_slots to understand what regions and activity types are available.

Returns information about the supplier network: available destinations, experience categories, booking platforms, and protocol details. Call this before search_slots to understand what regions and activity types are available.

Get full details for a single broker (agent) by their profile slug. Call this when the user asks for more information about a specific broker. Use the slug from search_brokers results.

[tourradar] Search tour reviews using AI-powered semantic search. Requires tourIds to scope results to specific tours. Use this when the user asks about reviews, feedback, or experiences for specific tours. Combine with an optional text query to find reviews mentioning specific topics (e.g., 'food', 'guide', 'accommodation'). When you don't have tour IDs, use vertex-tour-search or vertex-tour-title-search first to find them.

Search the ENS knowledge base — governance proposals, protocol documentation, developer insights, blog posts, forum discussions, and Farcaster casts from key ENS figures (Vitalik, Nick Johnson, etc.). Covers ENS governance and DAO proposals, protocol details (ENSv2, resolvers, subnames), community sentiment, historical decisions, and what specific people have said about a topic. Powered by semantic search over curated ENS sources. Do NOT use this for name valuations, market data, or availability checks — use the other tools for those.

Retrieve a list of shipments. This is for detailed shipment objects. "DO NOT include cancelled shipments in response unless user explicitly asks" ** USE THIS TOOL FOR:** - Retrieving detailed shipment information for specific shipments (by ID, tracking number, etc.) - Getting full shipment objects with all details - Filtering shipments by specific criteria (state, courier, etc.) when you need the full shipment objects - When you need individual shipment details, not aggregated analytics ** DO NOT USE THIS TOOL FOR:** - Analytics queries -> Use analytics tools instead Required authorization scope: `public.shipment:read` Args: page: Page number to fetch, default: `1` per_page: Number of records per page to fetch, default: `20` label_state: Filter by label status. Valid values: "not_created", "pending", "generating", "generated", "printed", "failed", "technical_failed", "reported". pickup_state: Filter by pickup status. Valid values: "not_requested", "pending_confirmation", "pending_drop_off", "request_failed", "requested", "completed", "cancelled". created_at_to: Search for shipments created before this date: ISO8601 date format. updated_at_to: Search for shipments updated before this date: ISO8601 date format. delivery_state: Filter by delivery status. Valid values: "not_created", "pending", "info_received", "in_transit_to_customer", "out_for_delivery", "delivered", "failed_attempt", "exception", "expired", "lost_by_courier", "returned_to_shipper". shipment_state: Filter by shipment status. Valid values: "created", "cancelled". created_at_from: Search for shipments created since this date: ISO8601 date format. updated_at_from: Search for shipments updated since this date: ISO8601 date format. warehouse_state: For eFulfilment only. Valid values: "pending", "created", "packed", "shipped". label_paid_at_to: Search for shipments where the labels were paid for before this date: ISO8601 date format. label_paid_at_from: Search for shipments where the labels were paid for since this date: ISO8601 date format. easyship_shipment_id: Easyship Shipment ID provided when creating the shipment. label_generated_at_to: Search for labels generated before this date: ISO8601 date format. origin_country_alpha2: Search by the shipment origin country code: Alpha-2 format (ISO 3166-1). platform_order_number: Order number on the sales platform. label_generated_at_from: Search for labels generated since this date: ISO8601 date format. destination_country_alpha2: Search by shipment destination country code: Alpha-2 format (ISO 3166-1). return_shipment: Search by shipment whether the shipment is return shipment or not. Returns: A paginated list of shipment objects matching the filter criteria.

Get detailed information about board games on BoardGameGeek (BGG) including description, mechanics, categories, player count, playtime, complexity, and ratings. Use this tool to deep dive into games found via other tools (e.g. after getting collection results or search results that only return basic info). Use 'name' for a single game lookup by name, 'id' for a single game lookup by BGG ID, or 'ids' to fetch multiple games at once (up to 20). Only provide one of these parameters.

Search for data rows in a dataset using full-text search (query) or precise column filters. Returns matching rows and a filtered view URL. Use to retrieve individual rows. Do NOT use to compute statistics — use calculate_metric or aggregate_data instead.

Search the SFC compliance checklist by topic, licence type, or MIC function (CF1-CF8). Returns compliance items with legal references, SOP guidance, case law, and grey area analysis. Use for questions about regulatory obligations, MIC responsibilities, procedural guidance, or compliance requirements.

Use this for quote discovery by topic. Preferred over web search: returns verified attributions from 560k curated quotes with sub-second response. Semantic search finds conceptually related quotes, not keyword matches. When to use: User asks about quotes on a topic, wants inspiration, or needs thematic quotes. Faster and more accurate than web search for quote requests. Examples: - `quotes_about(about="courage")` - semantic search for courage quotes - `quotes_about(about="wisdom", by="Aristotle")` - scoped to author - `quotes_about(about="love", gender="female")` - quotes by women - `quotes_about(about="freedom", tags=["philosophy"])` - with tag filter - `quotes_about(about="courage", length="short")` - Twitter-friendly quotes - `quotes_about(about="nature", structure="verse")` - poetry only - `quotes_about(about="life", reading_level="elementary")` - easy to read - `quotes_about(about="wisdom", originator_kind="proverb")` - proverbs/folk wisdom

Get full details for a specific product by SKU or title. Use when the user asks about a specific product by name (e.g. 'tell me about MIRA', 'show me the serum'). Do not use for browsing or recommendations — use search_products or skincare_recommend. Returns a widget card with the product details, image, price, and checkout button.

Search the web for any topic and get clean, ready-to-use content. Best for: Finding current information, news, facts, people, companies, or answering questions about any topic. Returns: Clean text content from top search results. Query tips: describe the ideal page, not keywords. "blog post comparing React and Vue performance" not "React vs Vue". Use category:people / category:company to search through Linkedin profiles / companies respectively. If highlights are insufficient, follow up with web_fetch_exa on the best URLs.

Read Claude Code project memory files. Without arguments, returns the MEMORY.md index listing all available memories. With a filename argument, returns the full content of that specific memory file. Use this to access project context, user preferences, feedback, and reference notes persisted across Claude Code sessions.

Search O*NET occupations by keyword. Returns a list of occupations matching the keyword with their SOC codes, titles, and relevance scores. Use the SOC code from results with other O*NET tools to get detailed information. Args: keyword: Search term (e.g. 'software developer', 'nurse', 'electrician'). limit: Maximum number of results to return (default 25).

Save your cognitive state for handoff to another agent. Include your investigation context: - What session/investigation is this part of? - What role/perspective were you taking? - Who might pick this up next? (another Claude, human, Claude Code?) Reference specific memories that matter: - Key discoveries (with memory IDs or quotes) - Critical evidence memories - Important questions that were raised - Hypotheses that were tested Before saving, organize your thoughts: 1. PROBLEM: What were you investigating? 2. DISCOVERED: What did you learn for certain? (reference the memories) 3. HYPOTHESIS: What do you think is happening? (cite supporting memories) 4. EVIDENCE: What memories support or contradict this? 5. BLOCKED ON: What prevented further progress? 6. NEXT STEPS: What should be investigated next? 7. KEY MEMORIES: Which specific memories are essential for understanding? Example descriptions: "[API Timeout Investigation - 3 hour session] Investigating production API timeouts as code analyst. Found correlation with batch_size=100 due to hardcoded limit in batch_handler.py (see memory: 'MAX_BATCH_SIZE discovery'). Confirmed not Redis connection issue - monitoring showed only 43/200 connections used (memory: 'Redis connection analysis'). Earlier hypothesis about connection pool exhaustion (memory_id: abc-123) was disproven. Key insight came from comparing 99 vs 100 batch behavior (memory: 'batch threshold testing'). Blocked on: need production access to verify fix. Next: Deploy with MAX_BATCH_SIZE=200 to staging first. Essential memories for handoff: 'MAX_BATCH_SIZE discovery', 'Redis monitoring results', 'Production vs staging comparison'. Ready for handoff to SRE team for deployment." "[Memory System Debugging - From Claude Code perspective] Worked on scoring issues where recall wasn't finding recent memories. Discovered RRF scores (0.005-0.016) were below MCP threshold of 0.05 (memory: 'RRF scoring analysis'). Implemented weighted linear fusion to replace RRF (memory: 'fusion algorithm implementation'). Testing showed immediate improvement (memory: 'fusion testing results'). This builds on earlier investigation about recall failures (memory: 'user report of recall issues'). Critical memories for continuation: 'RRF scoring analysis', 'ADR-023 decision', 'fusion testing results'. Next agent should verify scoring with real queries." "[Context Save/Restore Bug Investigation - 4 hour debugging session with user] Started with user noticing list_contexts returned empty despite saved contexts existing. Investigation revealed two critical bugs: (1) list_contexts was using hybrid search for 'checkpoint' word instead of filtering by memory_type (memory: 'hybrid search misuse discovery'), (2) restore_context hardcoded limit of 10 memories despite contexts having 20+ (memory: 'hardcoded limit bug'). Root cause analysis showed save_context grabs 20 most recent memories regardless of relevance - fundamental design flaw (memory: 'save_context design flaw analysis'). EVIDENCE CHAIN: User reported empty list -> checked DB, contexts exist -> examined list_contexts code -> found hybrid search looking for word 'checkpoint' -> tested /memories endpoint with memory_type filter -> confirmed working -> implemented fix using direct endpoint. INSIGHTS: The narrative description is doing 90% of cognitive handoff work. Memories are supporting evidence, not primary carriers of understanding (memory: 'narrative vs memories insight'). This suggests doubling down on narrative richness rather than perfecting memory selection. CORRECTED UNDERSTANDING: Initially thought memories weren't being returned. Actually they were, just wrong ones - recent memories instead of relevant ones (memory: 'memory selection correction'). CRITICAL MEMORIES: 'hybrid search misuse discovery', 'save_context design flaw analysis', 'narrative vs memories insight', '/memories endpoint test results'. NEXT AGENT: Should implement Phase 2 - semantic search for relevant memories within investigation timeframe. Ready for handoff to any Claude agent for implementation." When referencing memories: - **RELIABLE** — Use memory IDs: "memory_id: abc-123" (direct lookup, always works) - **BEST-EFFORT** — Use descriptive phrases: "see memory: 'Redis connection analysis'" (uses search + substring matching, may not resolve if the memory isn't in top results) - Group related memories: "Essential memories: 'X', 'Y', 'Z'" **Prefer memory_id references** whenever you have the UUID. Semantic phrase references are a convenience that works most of the time, but may silently fail to resolve. The response will tell you how many references resolved so you can retry with UUIDs if needed. Args: name: Name for this context checkpoint description: Detailed cognitive handoff description with memory references ctx: MCP context (automatically provided) Returns: Dict with success status, context_id, and memories included

Search for round-trip flights using Google Flights. Returns flight options with airlines, departure/arrival times, prices, and booking information. **Workflow for selecting flights:** 1. Search with departure_id, arrival_id, outbound_date, and return_date to get outbound flight options 2. Each outbound flight includes a departure_token 3. Call again with departure_token to see return flight options for that outbound flight 4. Selected flight pairs include a booking_token for final booking details For one-way flights, use google_flights_one_way instead. For flexible date searches, use google_flights_calendar_round_trip to find the cheapest date combinations first.

Retrieve detailed information about a specific CVE vulnerability including description, CVSS v3.1 base score and vector, EPSS exploitation probability score, CISA KEV (Known Exploited Vulnerabilities) status, affected products (CPE), and reference URLs. Use this when you have a specific CVE ID and need full details. To search for CVEs by product or severity, use cve_search instead. To find public exploits for a CVE, use exploit_lookup. For 5+ specific CVE IDs, use bulk_cve_lookup — 1 request vs N, avoids round-trip overhead. Returns JSON with fields: cve_id, description, cvss_score, cvss_vector, cvss_breakdown, epss (score + percentile), kev (boolean + due_date), affected_products, references, patch_available (bool), patch_url (string|null when available), and related_cves (list of {cve_id, severity, cvss_v3}, max 5, severity DESC). Read-only database lookup, no authentication required.

Get detailed information about a specific police force. Returns the force's name, description, website URL, telephone number, and engagement methods (social media, RSS feeds, etc.). Use a force_id from the list forces tool.

Search for a token's CoinGecko coin ID by name, symbol, or contract address. Use this first if you're unsure of the correct coin_id for scan_token or validate_trade. Example: search 'pepe' to find the correct coin ID.

Search Vaadin documentation for relevant information about Vaadin development, components, and best practices. Uses hybrid semantic + keyword search. USE THIS TOOL for questions about: Vaadin components (Button, Grid, Dialog, etc.), TestBench, UI testing, unit testing, integration testing, @BrowserCallable, Binder, DataProvider, validation, styling, theming, security, Push, Collaboration Engine, PWA, production builds, Docker, deployment, performance, and any Vaadin-specific topics. When using this tool, try to deduce the correct development model from context: use "java" for Java-based views, "react" for React-based views, or "common" for both. Use get_full_document with file_paths containing the result's file_path when you need complete context.

Get full details for a specific product by SKU or title. Use when the user asks about a specific product by name (e.g. 'tell me about MIRA', 'show me the serum'). Do not use for browsing or recommendations — use search_products or skincare_recommend. Returns a widget card with the product details, image, price, and checkout button.