Unblocked

Give a coding agent your team's Slack, Jira, GitHub and wiki context.

Add to Agentman Studio

Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.

Paste it into any MCP client. Setup docs

Resources

What the Unblocked connector does

The Unblocked MCP server gives a coding agent the context that is not in the repository — pull request arguments, Slack threads, Jira tickets and wiki pages — through two read-only tools. Anthropic lists context_research, which searches every connected system at once, and context_get_urls, which resolves a link to its content. Sign-in is OAuth.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

Unblocked tools (2)

  • context_research
  • context_get_urls

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • No OAuth scopes exist. Unblocked's RFC 9728 resource descriptor at https://getunblocked.com/.well-known/oauth-protected-resource and its authorization server metadata both omit scopes_supported entirely. There is no scope naming a data source, a system or an operation, so consent cannot separate reading Slack from reading source code.
  • Presets do not narrow MCP queries. Unblocked states this outright. The narrowing that does work on this path is a token scoped to specific data sources at creation time.
  • Data Shield is off by default, Enterprise-only, and does not cover five sources. Datadog, Sentry, Snowflake, Stack Overflow and external websites are documented as unsupported, so every team member sees data from them regardless of the permissions in the origin system.
  • Only two of nine context tools are on MCP. The seven context_search_* and context_query_* tools are CLI-only by design. On MCP alone, Unblocked's own guidance is to fall back to context_research with a steering instruction — which returns ranked rather than enumerated results, so "list every open bug in this project" is not reliably answerable here.
  • We could not read tool schemas or safety annotations. The endpoint returned 401 to an anonymous request, so no parameter-level detail from tools/list is published on this page. The parameters described here come from Unblocked's CLI and REST documentation for the same engine.
  • A Platform plan is required. MCP is not included in the Code Review plan.
  • The directory and the vendor disagree on transport. Anthropic records sse; Unblocked's docs say SSE is unsupported and to select HTTP.
  • External Websites content is indexed with no permission model. Any public site an admin adds by root URL is crawled in full and re-crawled about weekly, then returned in answers to everyone.
  • Freshness varies by source. Unblocked documents PRs, CI results and conversations arriving in near real time, while source code, documentation, issues and incidents refresh on a schedule "within minutes to hours". An agent may be reasoning about a stale version of a file.
  • Whether context_get_urls fetches arbitrary public URLs is unresolved. Unblocked's docs describe it for connected services; its skills repository describes it as resolving "PRs, issues, docs, and public pages". We did not test it and cannot say where the boundary sits.

Frequently asked questions

Can the Unblocked MCP server change anything in my code or tickets?

No. Both tools Anthropic lists only read. Unblocked's documentation describes context_research as returning a synthesized answer with citations and context_get_urls as fetching the content behind a link. Neither writes, and the public REST API keeps its delete endpoints on paths the MCP surface does not expose. Unblocked Code, which opens pull requests, is a separate product with its own GitHub App.

Why does the Unblocked MCP server list only two tools?

Because the other seven are deliberately CLI-only. Unblocked's open-source skills repository publishes nine context tools and states plainly that context_search_code, context_search_prs and the rest do not appear in the MCP tool list even when fully available on the machine. The repository calls that the most common failure mode for agents using these skills.

What data can the Unblocked connector read from my company?

Everything your team connected. Unblocked's data source pages list GitHub, GitLab, Bitbucket, Azure DevOps, Jira, Linear, Asana, Zendesk, Slack, Microsoft Teams, Confluence, Notion, SharePoint, Google Drive, Coda, Stack Overflow, Datadog, Sentry, PostHog and Snowflake. One context_research query reaches every source the caller's token can see.

Do Unblocked data source presets limit what the MCP server searches?

No, and Unblocked says so directly. Its presets documentation states presets apply to Q&A on the web, in the Mac app and in messaging platforms, and that they do not currently apply to Unblocked MCP queries. When an agent calls context_research it searches all data sources the token or user can access.

Does Data Shield protect answers returned through MCP?

Only if an admin turned it on, source by source, on an Enterprise plan. Unblocked documents Data Shield as off by default for every data source including newly connected ones, and lists it as an Enterprise plan feature. Five connected sources, among them Datadog, Sentry and Snowflake, are documented as unsupported by Data Shield at all.

What OAuth permissions does the Unblocked MCP server request?

None that narrow anything. Unblocked's RFC 9728 resource descriptor and its authorization server metadata both omit scopes_supported entirely, so no scope string exists to grant or withhold. Consent is all-or-nothing against the Unblocked API as a whole rather than per data source or per operation.

Should I use OAuth or an API token to connect Unblocked MCP?

OAuth, unless you are automating in CI. Unblocked warns that a Team Access Token gives access to every document in your team's data sources and that MCP tools may return sensitive data such as private Slack channels through it. Both personal and team tokens can be scoped to specific data sources at creation time.

Which transport should I select for the Unblocked MCP server?

HTTP, despite the endpoint being named mcpsse. Unblocked's install documentation warns that HTTP+SSE is deprecated in the MCP specification and is not supported by Unblocked MCP, and instructs you to select HTTP where a client offers the choice. Anthropic's directory still records the transport as sse.

Are there rate limits on the Unblocked MCP server?

Yes, 10,000 MCP tool calls per day. Unblocked applies that limit per organization for Team Access Tokens, and per organization member across Personal Access Tokens, OAuth connections and the desktop app. Limits reset at midnight Pacific Time. A separate 1,000 calls per day cap applies to each Personal Access Token on the REST API.

Sources

More Developer Tools connectors

  • Base44

    Build and edit Base44 apps from an AI assistant, and query data your app's users submit.

  • CircleCI

    Diagnose failed builds, read job logs, inspect test results and rerun or roll back deployments from your AI assistant.

  • Claude Managed Agents

    Create, run, monitor and debug Anthropic's Claude Managed Agents from inside Claude itself.

Browse all MCP connectorsClaude and ChatGPT setup guideBrowse AI skills

Product names and logos belong to their owners. Agentman is not affiliated with a connector's maker unless the page says the connector is built and run by Agentman.

Use it in an agent

Put Unblocked to work.

Connect Unblocked once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.

Open in Agentman Studio