Atlas / MCP servers / lanes-sh / link

linkCAUTION

Name: link
Author: lanes-sh

mcp/lanes-sh/link

MCP server by lanes-sh · chatgpt · claude · mcp · Audited 2026-10-08

⚑ Report / re-scan Source ↗

Lanes Link is a private, self-hostable MCP that connects your accounts, memory and skills to every AI agent you use.

Verdict

CAUTION

Grade

Trust score

38 /100

Exposed tools

125 107r · 10w · 8d

Transport

stdio · streamable-http

License

Apache-2.0

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

[](https://www.npmjs.com/package/@lanes-sh/link) [](LICENSE) [](https://github.com/lanes-sh/link/actions/workflows/ci.yml)

One endpoint you own, holding everything your agents need to know you: connections, memory, tasks, files, contacts, and secrets.

Connect your mail, calendar, files and notes once. Every agent you use, Claude, ChatGPT, and anything else that speaks MCP, reaches all of it through that one endpoint. Change your AI and you keep your context, because none of it ever lived in the agent.

Quickstart

Needs Bun 1.3.11+ and a Lanes sign-in.

$ bun install -g @lanes-sh/link
$ lanes auth login
$ lanes link profile add personal
$ lanes link start
ok    serving http://127.0.0.1:7337/mcp
profiles: personal

Leave that running. In a second shell, point every agent you have installed at it:

$ lanes link mcp add
ok    registered lanes-link with Claude Code (user scope)
ok    registered lanes-link with Codex

Your memory, tasks, files, skills, entities and vault work now, with nothing to authorise. Mail and calendar are the next step: lanes link connect gmail.

[Full quickstart](https://lanes.sh/docs/link/quickstart)

Why

Connect once, use everywhere. No per-agent integrations, no re-authorising every new tool.
You decide what agents can touch. Permi

Read from source at commit 0bff5f253acbOBSERVED · 2026-10-08

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code

claude mcp add link -- npx -y @lanes-sh/[email protected]

claude-desktop

{
  "mcpServers": {
    "link": {
      "command": "npx",
      "args": [
        "-y",
        "@lanes-sh/[email protected]"
      ]
    }
  }
}

Exposed tools (125)

107 read · 10 write · 8 destructive. Blast radius: 8 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`Airtable`	read	Bases, tables, records, fields, and schema, via Airtable\
`Algolia`	read	Search indices, records, queries, and synonyms, via Algolia\
`Amplitude`	read	Events, charts, cohorts, and user activity, via Amplitude\
`Apify`	read	Actors, runs, datasets, and scraped results, via Apify\
`Asana`	read	Tasks, projects, portfolios, and workspaces, via Asana\
`Atlassian`	read	Jira issues, Confluence pages, and Compass components, via Atlassian\
`Attio`	read	Records, lists, notes, and tasks in the CRM, via Attio\
`Box`	read	Files, folders, and metadata in Box, via Box\
`Buildkite`	read	Pipelines, builds, jobs, and artifacts, via Buildkite\
`Calendly`	read	Scheduled events, invitees, event types, and availability, via Calendly\
`Canva`	read	Designs, folders, brand templates, assets, and exports, via Canva\
`CircleCI`	read	Pipelines, workflows, jobs, and test results, via CircleCI\
`ClickUp`	read	Tasks, lists, spaces, docs, and time entries, via ClickUp\
`Close`	read	Leads, contacts, opportunities, and activities in the CRM, via Close\
`Contentful`	read	Entries, assets, content types, and spaces, via Contentful\
`Datadog`	read	Metrics, logs, monitors, incidents, and dashboards, via Datadog\
`Discord`	write	Post announcements, read channels, and triage messages in the servers your bot has been added to, via the Discord v10 HTTP API.
`Dropbox`	read	Files, folders, shared links, and file requests, via Dropbox\
`Expensify`	read	Expenses, reports, and receipts, via Expensify\
`Figma`	read	Files, designs, components, and Dev Mode context, via Figma\
`Fireflies`	read	Meeting transcripts, summaries, and action items, via Fireflies\
`Flagsmith`	read	Feature flags, segments, and environments, via Flagsmith\
`Gamma`	read	Presentations and documents, generated and read back, via Gamma\
`GitHub`	read	Repositories, issues, pull requests, and workflow runs, via GitHub\
`Gmail`	read
`Grafana`	read	Dashboards, datasources, queries, and alert rules, via Grafana\
`Heroku`	write	Apps, dynos, add-ons, releases, and logs, via Heroku\
`HubSpot`	read	CRM contacts, companies, deals, and engagements, via HubSpot\
`Hygraph`	read	Content entries, models, and schema, via Hygraph\
`Insightly`	read	Contacts, organisations, opportunities, and projects, via Insightly\
`Jam`	read	Bug reports, with their console logs, network calls, and repro steps, via Jam\
`Klaviyo`	read	Profiles, lists, segments, campaigns, and flows, via Klaviyo\
`Linear`	read	Issues, projects, comments, and cycles, via Linear\
`Mercury`	read	Accounts, balances, transactions, and cards, via Mercury\
`Miro`	read	Boards, frames, sticky notes, and shapes, via Miro\
`Mixpanel`	read	Events, funnels, retention, and cohorts, via Mixpanel\
`Mux`	read	Video assets, live streams, and playback analytics, via Mux\
`Navan`	read	Trips, bookings, and travel expenses, via Navan\
`Neon`	read	Postgres projects, branches, SQL, and docs, via Neon\
`Netlify`	read	Sites, deploys, functions, and environment variables, via Netlify\
`Notion`	read	Pages, databases, comments, and workspace search, via Notion\
`OneDrive`	read	Browse, search, read, and organise files in OneDrive, via Microsoft Graph.
`Paddle`	read	Products, prices, subscriptions, and transactions, via Paddle\
`PayPal`	read	Invoices, orders, payments, subscriptions, and disputes, via PayPal\
`PostHog`	read	Events, insights, feature flags, and session replays, via PostHog\
`Prisma`	read	Postgres databases, schema, and migrations, via Prisma\
`Ramp`	read	Cards, transactions, reimbursements, and spend limits, via Ramp\
`Recurly`	read	Subscriptions, invoices, and accounts, via Recurly\
`Reddit`	write	Read subreddits, posts, comments, and search, and post, comment, vote, and edit as your account, via the Reddit API.
`Remote`	read	Employees, contracts, payroll, and time off, via Remote\
`Render`	read	Services, deploys, logs, and environment variables, via Render\
`Replicate`	read	Models, predictions, and deployments, via Replicate\
`Resend`	read	Transactional email, domains, and delivery events, via Resend\
`Riverside`	read	Recordings, transcripts, and clips, via Riverside\
`Rootly`	read	Incidents, alerts, retrospectives, and on-call schedules, via Rootly\
`RudderStack`	read	Sources, destinations, and event streams, via RudderStack\
`Salesloft`	read	Cadences, people, and sales activity, via Salesloft\
`Sanity`	read	Documents, datasets, schema, and content releases, via Sanity\
`Sentry`	read	Issues, events, stack traces, and releases, via Sentry\
`Shared`	read
`Shortcut`	read	Stories, epics, iterations, and workflows, via Shortcut\
`Slack`	read	Messages, threads, channels, files, and canvases, via Slack\
`Square`	read	Payments, orders, catalog, inventory, and customers, via Square\
`Storyblok`	read	Stories, components, assets, and spaces, via Storyblok\
`Stripe`	read	Payments, customers, invoices, subscriptions, and balances, via Stripe\
`Supabase`	read	Projects, database schema, SQL, edge functions, and docs, via Supabase\
`Tavily`	read	Web search and page content extraction, via Tavily\
`Todoist`	read	Tasks, projects, sections, labels, and filters, via Todoist\
`Vercel`	read	Projects, deployments, build logs, and domains, via Vercel\
`Vimeo`	read	Videos, folders, showcases, and analytics, via Vimeo\
`Webflow`	read	Sites, pages, CMS collections, and items, via Webflow\
`Whimsical`	read	Boards, flowcharts, wireframes, and mind maps, via Whimsical\
`Wix`	read	Sites, stores, bookings, and CMS data, via Wix\
`Workable`	read	Jobs, candidates, and interviews, via Workable\
`Zapier`	read	Zaps, and the actions they reach across thousands of apps, via Zapier\
`add`	write	Record something to be done. This is where
`author`	destructive	Read, create, and delete the skills themselves.
`boom`	read	always throws
`bunq`	read	Bank accounts, balances, transaction history, and payments — including batches and drafts that wait for approval in the bunq app.
`compose`	read	Something the vendor document cannot describe.
`delete_note`	destructive	Remove a note from this connection.
`diff`	read	The unified diff
`doc`	read	a document
`echo`	read	echo
`entity`	read	One declared entity, addressed by its id.
`entry`	read	One stored memory entry, addressed by its id.
`file`	read	One stored file, addressed by its name. Text comes back as text; anything else is described rather than encoded.
`find`	read	Find entities by name, alias, address, type, tag, attribute or relationship. All criteria
`forget`	destructive	Delete an entity. Edges pointing at it from other entities are left alone and are reported,
`get`	read	Return a text file\
`get_note`	read	read a note
`grant_policy`	read	widen my own permissions
`link`	write	Add one edge to an entity, leaving everything else on it alone. The edge is written only on
`list`	read	Every file kept in this profile, newest first, with its type and size. This is the whole index — an asset carries no description, so what a file is for belongs in memory.
`list_notes`	read	List the keys of every note stored against this connection.
`list_things`	read	From the document.
`manage.get`	write	Return a skill exactly as stored, frontmatter included — what to edit before writing it back.
`manage.list`	read	Every skill that exists, with its description and arguments. The prompt list shows only the ones policy permits; this shows what is stored.
`manage.remove`	destructive	Remove a skill and the prompt it provided.
`manage.write`	write	Create or replace a skill. The text is a whole Markdown document: YAML frontmatter carrying
`monday.com`	read	Boards, items, groups, columns, and updates, via monday.com\
`note`	read	A stored note, addressed by key.
`notes.get`	read	One note, by identifier.
`notes.list`	read	Identifiers of the notes in this account, and a token for the next page.
`overview`	read	List the accounts reachable in this profile and the providers that could be connected.
`procedure`	read	a reusable procedure
`provider`	read	The console steps, the values needed, and the exact command that connects it.
`purge`	destructive	destructive
`put`	write	Store or replace a secret under an id. The item becomes readable only after the endpoint restarts, and only if policy grants
`read`	read	List and read stored files.
`remove`	destructive	Remove a file and its bytes. There is no trash.
`review-diff`	read	Review a diff for correctness
`round_trip`	read	stage a file and name it back
`search`	read	Find entries whose title, tags, or body contain the query. Case-insensitive substring matching, not ranked relevance.
`set_note`	write	store a note
`stage`	read	Hold a file for a later call, anywhere in this profile. Name one source — data for a file
`store`	read	Keep a file in this profile, permanently and by name. Name exactly one source — path, url, handle, asset, or data — and the endpoint reads the bytes itself. Storing under a name that exists replaces it. A stored file attaches to anything afterwards as {
`style`	read	style
`subject`	read	what about
`task`	read	One task, addressed by its id.
`unbounded`	read	a resource space too large to enumerate
`update`	write	Change a task in place — most often its status. Marking something done is an update, not a delete: the record of having done it is the useful part. Omitted fields are left as they are.
`users.drafts.delete`	destructive
`users.messages.list`	read
`write`	destructive	Store and delete files.

Trust audit

CAUTIONgrade F · trust 38/100 Install with care. The audit found things worth knowing before you trust its output.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	WARN
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED