attio-mcp
Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.
Needs to start: nothing
Returns a result when called with no setup: 0 of 3 tools tried
Contacts while installing and starting: nothing
What it contacted
Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.
Starting it
npx -y -p attio-mcp@1.7.0 attio-mcp
Installing a package runs its code. Use a sandbox for anything you have not read.
Calling its tools
3 tools that take no input and look read-only were called once with nothing. 0 returned a result.
aaa-health-check | No usable answer replied, but asks for setup: { "ok": true, "name": "attio-mcp", "timestamp": "2026-11-16T11:32:47.751Z", "environment": "production", "needs_api_key": |
get-lists | No usable answer tool reported an error: ERROR [unknown_error]: Invalid API key: API key must be a non-empty string (source: provided configuration) Details: {"method":"GET","u |
list-workspace-members | No usable answer tool reported an error: Authentication failed. Please check your credentials. [Dev Info: Could not initialize Attio client: API client not initialized and no va |
During the calls it contacted nothing it had not already contacted.
Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.
What its tools tell an agent
In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.
Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.
Tools
aaa-health-check
Run a lightweight health probe that echoes deployment metadata. Never query Attio APIs, mutate data, or require credentials. Accepts optional echo text; returns JSON payload as text for MCP clients. If unavailable, review Smithery sandbox logs or restart the server process.
smithery_debug_config
Retrieve non-sensitive diagnostic information about Smithery runtime configuration propagation. Never expose credentials or auth state, write data, or modify configuration. Returns runtime, workspace, and context-storage diagnostics only. Read-only operation. Use this tool to compare runtime mode, workspace configuration, and context storage state across deployments.
search_records resource_type, query, filters, search_type, fields, match_type, sort, date_from, date_to, created_after, created_before, updated_after, updated_before, timeframe, date_field, limit, offset
Search across companies, people, deals, tasks, and records. Never create or modify records. Returns max 100 results (default: 10). use discover_record_attributes to find searchable fields.
get_record_details resource_type, record_id, fields
Fetch a single record with enriched attribute formatting. Never search or filter result sets; use search_records tools instead. Requires resource_type and record_id; optional fields filter output. Validate record IDs with search_records before retrying.
create_company record_data, return_details
Create one company in Attio without choosing resource_type. Never update existing companies or create deals, people, or tasks. May require explicit user approval from the host. Requires record_data with company attribute API slugs such as name, domains, or website. If fields are rejected, call discover_record_attributes with resource_type companies.
39 more tools
update_company record_id, record_data, return_details
Update one company in Attio without choosing resource_type. Never create records, delete records, or update deals. May require explicit user approval from the host. Requires record_id and record_data with company attribute API slugs. Call get_record_details or search_records for companies first to confirm the target company ID.
create_deal record_data, return_details
Create one deal in Attio without choosing resource_type. Never update existing deals or create companies, people, or tasks. May require explicit user approval from the host. Requires record_data with deal attribute API slugs such as name, stage, value, owner, or associated company/person references. If stage or owner values are rejected, call discover_record_attributes or get_record_attribute_options for deals.
update_deal record_id, record_data, return_details
Update one deal in Attio without choosing resource_type. Never create records, delete records, or update companies. May require explicit user approval from the host. Requires record_id and record_data with deal attribute API slugs. Call search_records for deals first to confirm the target deal ID and current stage.
create_record resource_type, record_data, return_details
Create new Attio records across supported resource types when a scoped tool is not available. Prefer create_company or create_deal for common company/deal writes. Never update existing records, attach files, or bypass required fields. May require explicit user approval from the host. Requires resource_type plus record_data that matches discover_record_attributes output. If validation fails, call discover_record_attributes to confirm required fields and enums. If a select/status value is rejected, call get_record_attribute_options for that attribute to list valid options before retrying.
update_record resource_type, record_id, record_data, return_details
Update existing Attio record fields across supported resource types when a scoped tool is not available. Prefer update_company or update_deal for common company/deal writes. Never create new records, delete data, or manage list memberships. May require explicit user approval from the host. Requires resource_type, record_id, and record_data; supports partial updates with schema validation. Call get_record_details first to inspect the latest values before editing. If a select/status value is rejected, call get_record_attribute_options for that attribute to list valid options.
delete_record resource_type, record_id
Delete an Attio record from its object (company, person, deal, task). Never cascade delete related data or clean up list memberships automatically. May require explicit user approval from the host. Requires record_id and resource_type; operation is irreversible once confirmed. If uncertain, fetch with get_record_details to confirm the target before deletion.
get_record_attributes resource_type, record_id, categories, fields
Retrieve attribute metadata for a given resource type. Never modify schema definitions or record data. Requires resource_type; optional categories narrows groups. Use discover_record_attributes for grouped schema discovery.
discover_record_attributes resource_type, categories
Discover available attributes (standard/custom) for a resource. Never alter schema or create fields. Requires resource_type; optional categories selects subsets. For select/status attributes, use get_record_attribute_options to fetch valid values before creating or updating records.
get_record_attribute_options resource_type, attribute, show_archived
Get valid options for select, multi-select, and status attributes to avoid "Cannot find select option" errors. Never return options for text, number, or other non-option types. Requires resource_type and attribute slug/ID. Use discover_record_attributes to find option-based attributes first. Use retrieved option titles when calling create-record or update-record.
get_record_info resource_type, record_id
Retrieve enriched info subsets (contact, business, social) for a record. Never search lists of records or mutate data. Requires resource_type, record_id, and info_type (contact|business|social). Use get_record_details if you need the full record payload.
create_note resource_type, record_id, title, content, format
Create note for companies, people, or deals with full markdown support. Never update or delete notes; creates only. May require explicit user approval from the host. Requires resource_type, record_id, title, content. Set format="markdown" for rich formatting: headings (# ## ###), lists (- or 1.), nested bullets (2-space indent), bold (**text**), code blocks. Use \n for line breaks. If record not found, use search_records first.
list_notes resource_type, record_id, parent_record_id, limit, offset
Retrieve notes for a record with timestamps. Never create or modify notes; read-only. Requires resource_type, record_id; sorted by creation date. If empty, verify record has notes with get_record_details.
get_record_interactions resource_type, record_id
Fetch interaction metadata (first/last email, calendar, interaction timestamps and owners) for a person or company record. Never retrieve full email content or activity feeds; Attio only exposes interaction timestamps and owner actors. Requires resource_type (people or companies) and record_id. Returns system-generated interaction attributes. If no interactions found, verify the record exists with get_record_details. For activity content, use search_records_by_content with content_type=activity.
merge_records resource_type, record_id, secondary_record_id, dry_run, confirm, keep_from_leftover, skip_leftover_attributes, override_linked_mismatch, plan_fingerprint
Preview and, after explicit confirmation, safely merge two Attio deal records using the native beta merge operation. Never merge people or companies, merge more than one pair, or retry an indeterminate native merge. May require explicit user approval from the host. Dry-run defaults to true. Execute requires dry_run=false, confirm=true, and plan_fingerprint from that dry-run. For HTTP 202, wait and call get_record_details with new_record_id later; both original ids are unreadable after native merge.
search_records_advanced resource_type, query, filters, sort_by, sort_order, limit, offset
Search companies, people, deals, or tasks with complex nested filters (e.g., find deals by owner+stage, companies by industry+location). Never mutate records; use update_record or delete_record. Supports filter groups, scoring, pagination, and up to 100 items. Requires resource_type parameter. If filters fail, fetch valid attributes via discover_record_attributes.
search_records_by_relationship relationship_type, source_id, target_resource_type, listId, limit, offset
Search records using relationship anchors (list, company, people). Never modify memberships; use list tools for writes. Requires resource_type and related resource identifier. Use search_records to resolve IDs before calling.
search_records_by_content resource_type, content_type, search_query, limit, offset
Search record content (notes, activity, communications). Never modify note content or attachments. Requires resource_type and content_query; optional fields array. Narrow scope with fields or switch to search_records_advanced.
search_records_by_timeframe resource_type, timeframe_type, start_date, end_date, date_field, limit, offset
Filter records by creation, update, or interaction timeframes. Never modify lifecycle state or scheduling follow-ups. Requires resource_type; provide timeframe or explicit date boundaries. Call search_records if timeframe filters are too restrictive.
batch_records resource_type, operations, operation_type, records, record_ids, limit, offset
Execute batched record operations (create/update/delete/get/search). Never ignore approval guardrails; hosts may require confirmation. May require explicit user approval from the host. Use scoped single-record tools for one company or deal write. operation_type must be specified for legacy payloads; operations arrays must use explicit create/update/delete entries. Run search_records first to stage IDs or payloads for batching.
batch_search_records resource_type, queries, limit, offset
Execute multiple searches in parallel and return grouped results. Never mutate or import data; use batch_records for write operations. Provide queries array (1–10 items recommended) and resource_type. If queries fail, retry individually using search_records.
search query, type, limit
Run lightweight compatibility search across companies, people, lists, and tasks for ChatGPT MCP. Never support complex filters, batch queries, or return rich record payloads (use search_records* tools). Requires query string (min 1 char); optional type filter; limit up to 25 results per call. When you need pagination or attribute filtering, switch to search_records or search_records_advanced.
fetch id
Return the canonical Attio record payload for a search connector reference. Never perform write operations or resolve IDs not emitted by the search tool. Accepts identifier generated by search (<resource>:<uuid>) and returns JSON serialized as text. If fetch fails, rerun search to refresh the identifier or verify the record still exists.
get-lists
Retrieve all CRM lists (sales pipelines, lead stages, customer segments). Never create or modify lists, only reads existing lists. Returns all lists visible to the authenticated workspace. Use get-list-details to inspect individual list schemas.
get-record-list-memberships recordId, objectType, includeEntryValues, batchSize
Find all lists containing a specific company or person record. Never modify list memberships or retrieve list entries. Requires recordId; processes 5 lists in parallel by default (max 20). If record not found, verify recordId with search_records first.
get-list-details listId
Retrieve schema and configuration for a specific list (stages, fields, attributes). Never modify list structure or retrieve list entries. Requires valid list UUID or slug; accepts both formats. Use get-lists to discover available list IDs and slugs first.
get-list-entries listId, limit, offset
Retrieve all records in a list with pagination (companies, people in pipelines). Never filter entries or modify list memberships. Requires list UUID (not slug); default limit 20, max per page varies by API. Use filter-list-entries for attribute-based filtering instead.
filter-list-entries listId, attributeSlug, condition, value, filters, parentObjectType, parentAttributeSlug, parentRecordId, limit, offset
Filter list entries with flexible parameter modes. Auto-detects mode based on parameters provided.
**Mode 1 - Simple Filtering (Single Attribute):**
Filter entries by a single attribute condition.
Required parameters: listId, attributeSlug, condition, value
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"attributeSlug": "status",
"condition": "equals",
"value": "active"
}
**Mode 2 - Advanced Filtering (Multi-Condition):**
Filter entries with multiple conditions using AND/OR logic.
Required parameters: listId, filters (object with filters array and matchAny flag)
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"filters": {
"filters": [
{"attribute": {"slug": "status"}, "condition": "equals", "value": "active"},
{"attribute": {"slug": "priority"}, "condition": "greater_than", "value": 5}
],
"matchAny": false
}
}
**Mode 3 - Parent Attribute Filtering:**
Filter entries by parent record attributes (e.g., company industry, person role).
Required parameters: listId, parentObjectType, parentAttributeSlug, condition, value
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"parentObjectType": "companies",
"parentAttributeSlug": "categories",
"condition": "contains",
"value": "Technology"
}
**Mode 4 - Parent UUID Filtering:**
Filter entries by exact parent record UUID (convenience mode, fastest).
Required parameters: listId, parentRecordId
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"parentRecordId": "660e8400-e29b-41d4-a716-446655440001"
}
**Mode Detection:**
The tool automatically detects which mode to use based on the parameters you provide.
You must provide parameters for exactly ONE mode per call.
**Pagination:**
All modes support optional 'limit' and 'offset' parameters for pagination.
**Migration Guide:**
- Replaces: advanced-filter-list-entries → Use Mode 2 (filters parameter)
- Replaces: filter-list-entries-by-parent → Use Mode 3 (parentObjectType + parentAttributeSlug)
- Replaces: filter-list-entries-by-parent-id → Use Mode 4 (parentRecordId)advanced-filter-list-entries listId, filters, limit, offset
[DEPRECATED] This tool has been consolidated into 'filter-list-entries'. Please use 'filter-list-entries' with Mode 2 (Advanced) parameters instead: - Pass a 'filters' object with 'filters' array and 'matchAny' flag - All functionality remains identical This tool will be removed in version 2.0.0. --- Filter entries with multi-condition queries (AND/OR logic). Never modify entries; read-only. Requires listId, filters array; matchAny for OR logic. Use filter-list-entries for single conditions.
add-record-to-list listId, recordId, objectType, initialValues
[DEPRECATED] This tool has been consolidated into 'manage-list-entry'. Please use 'manage-list-entry' with Mode 1 (Add) parameters instead: - Pass 'recordId', 'objectType', and optionally 'initialValues' - All functionality remains identical This tool will be removed in version 2.0.0. Add company or person to list with optional initial values. Never create records; record must exist first. May require explicit user approval from the host. Requires list UUID, record UUID, object type. If not found, create record first with create-record.
remove-record-from-list listId, entryId
[DEPRECATED] This tool has been consolidated into 'manage-list-entry'. Please use 'manage-list-entry' with Mode 2 (Remove) parameters instead: - Pass 'entryId' only (do not include 'attributes' or 'recordId') - All functionality remains identical This tool will be removed in version 2.0.0. Remove company or person from list (membership only). Never delete underlying record; membership only. May require explicit user approval from the host. Requires list UUID, entry UUID (not record UUID). Use get-list-entries to find entry UUID.
update-list-entry listId, entryId, attributes
[DEPRECATED] This tool has been consolidated into 'manage-list-entry'. Please use 'manage-list-entry' with Mode 3 (Update) parameters instead: - Pass 'entryId' and 'attributes' parameters - All functionality remains identical This tool will be removed in version 2.0.0. Update list entry attributes (stage, status, custom fields). Never update record attributes; use update-record for that. May require explicit user approval from the host. Requires list UUID, entry UUID, attributes object. Use get-list-details for valid attributes and values.
manage-list-entry listId, recordId, objectType, initialValues, entryId, attributes
Manage list entries with flexible action modes. Auto-detects action based on parameters provided.
**Mode 1 - Add Entry (Add record to list):**
Add a company or person record to a list with optional initial values.
Required parameters: listId, recordId, objectType
Optional parameters: initialValues
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"recordId": "660e8400-e29b-41d4-a716-446655440001",
"objectType": "companies",
"initialValues": {"stage": "Prospect"}
}
**Mode 2 - Remove Entry:**
Remove an entry from a list.
Required parameters: listId, entryId
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"entryId": "770e8400-e29b-41d4-a716-446655440002"
}
**Mode 3 - Update Entry:**
Update attributes on an existing list entry.
Required parameters: listId, entryId, attributes
Example:
{
"listId": "550e8400-e29b-41d4-a716-446655440000",
"entryId": "770e8400-e29b-41d4-a716-446655440002",
"attributes": {"stage": "Qualified"}
}
**Mode Detection:**
The tool automatically detects which action to perform based on the parameters you provide.
You must provide parameters for exactly ONE mode per call.
**Migration Guide:**
- Replaces: add-record-to-list → Use Mode 1 (recordId + objectType)
- Replaces: remove-record-from-list → Use Mode 2 (entryId only)
- Replaces: update-list-entry → Use Mode 3 (entryId + attributes)filter-list-entries-by-parent listId, parentObjectType, parentAttributeSlug, condition, value, limit, offset
[DEPRECATED] This tool has been consolidated into 'filter-list-entries'. Please use 'filter-list-entries' with Mode 3 (Parent Attribute) parameters instead: - Pass 'parentObjectType', 'parentAttributeSlug', 'condition', and 'value' - All functionality remains identical This tool will be removed in version 2.0.0. --- Filter entries by parent record attributes (industry, role). Never search multiple lists or modify records. Requires listId, parentObjectType, parentAttributeSlug, condition, value. Use discover_record_attributes for valid slugs.
filter-list-entries-by-parent-id listId, recordId, limit, offset
[DEPRECATED] This tool has been consolidated into 'filter-list-entries'. Please use 'filter-list-entries' with Mode 4 (Parent UUID) parameters instead: - Pass 'parentRecordId' parameter - Note: Parameter renamed from 'recordId' to 'parentRecordId' for consistency - All functionality remains identical (fastest filtering mode) This tool will be removed in version 2.0.0. --- Filter entries by exact parent record UUID. Never search multiple lists. Requires list UUID, record UUID; faster than attribute filtering. Use get-record-list-memberships for workspace-wide search.
create-list name, parent_object, description, template, attributes, dry_run
Create a new CRM list (sales pipeline, recruiting tracker, support queue). Supports templates for quick setup with smart defaults. Never update existing lists or manage list entries. Requires name and parent_object. Template expansion fills defaults before validation. Dry-run mode previews without creating. Use get-lists to verify the list was created. Use update-list-configuration to modify after creation.
update-list-configuration listId, attributes, dry_run
Update configuration of an existing list (name, description, custom fields). Detects immutable fields and provides clear errors. Never change parent_object (immutable) or manage list entries. Requires listId. Immutable fields (parent_object) are rejected before the API call. Dry-run mode previews changes. Use get-list-details to inspect current configuration. Use create-list for a different parent_object.
list-workspace-members search, page, pageSize
List workspace members to plan assignments and access checks. Never change access levels or invite new members; read-only. Supports optional search, pagination (1-100 per page, default 25). Use search-workspace-members for targeted lookups.
search-workspace-members query
Search workspace members by name, email, or access role. Never modify member profiles or permissions; lookup only. Requires query string (minimum 2 characters). If no results, list-workspace-members provides the full roster.
get-workspace-member memberId
Retrieve profile and access details for one workspace member. Never update member information or change permissions. Requires workspace_member_id from list/search results; read-only. Use list-workspace-members to confirm the memberId before retrying.
Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.
Install facts
| License | Apache-2.0 |
|---|---|
| Needs | node >=20.0.0 |
| Source | github.com/kesslerio/attio-mcp-server |
| Install scripts | attio-mcp declare them. This runner does not run them. |
| Tarball | sha512-F8LAvs06mluXokFxnEX5uCtcubsoo/ci4J8JBLkdJq3mu2A1IG+FLIC1Jw+E866/vsGFRBnOM/JH2xk3KxXFxQ== |
Check this report yourself
This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.
Report as JSONExact text and tokenHow to verify it
Show it on your README
[](https://bragproof.smaverk.com/r/attio-mcp)
New version out? Ask for a fresh report.