Stride — AI software delivery platform
RemoteAbout Stride — AI software delivery platform
Stride plugins let coding agents work through a team's Stride backlog, from picking up a specified story to returning completed work for review. They provide a two-command workflow for Claude Code, Codex and Cursor, with browser OAuth sign-in.
Key features
- 01Pick up assigned or staged stories
- 02Load specs, designs and related records
- 03Set story status and branch name
- 04Check tests and acceptance criteria
- 05Return completed stories for review
- 06Support Claude Code, Codex and Cursor
Use cases
- 01Start work on a staged backlog story
- 02Select a specific story by its key
- 03Verify changes against acceptance criteria
- 04Hand finished work back for review
MCP support
Each plugin connects the coding agent to Stride's MCP server at `https://www.stride.page/api/mcp/stride`. Users authenticate through browser OAuth; headless runs can use an agent key.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://www.stride.page/api/mcp/strideclaude mcp add --transport http page-stride-stride https://www.stride.page/api/mcp/stride{"mcpServers":{"page-stride-stride":{"type":"http","url":"https://www.stride.page/api/mcp/stride"}}}Open in Cursor{"servers":{"page-stride-stride":{"type":"http","url":"https://www.stride.page/api/mcp/stride"}}}Open in VS Code{"mcpServers":{"page-stride-stride":{"type":"http","url":"https://www.stride.page/api/mcp/stride"}}}{"mcpServers":{"page-stride-stride":{"serverUrl":"https://www.stride.page/api/mcp/stride"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEStride plugins for Claude Code, Codex and Cursor
Work your Stride backlog from your coding agent. Two commands do the loop:
| Claude Code | Codex | Cursor | |
|---|---|---|---|
| Pick up your next story, with its full spec, and start it | /stride:next | $stride:next | /stride-next |
| Check it against its acceptance criteria and hand it back for review | /stride:done | $stride:done | /stride-done |
Next pulls the story a teammate staged for you with "Hand to your agent"
or, when nothing is staged, the top story assigned to you that is still to do
(the active sprint first, then by priority). You get its acceptance criteria,
linked designs and documents, relations and open defects in one call. It moves
the story to In Progress and sets up a branch named after its key. Name a key
(/stride:next ORB-526) to pick up a particular story.
Done runs your tests, checks every acceptance criterion against the change, and moves the story to In Review with a comment saying what changed. If a criterion is not met, it says what remains and leaves the status alone. It asks before committing, and pushes only when you ask.
Every plugin connects your agent to Stride's MCP server,
https://www.stride.page/api/mcp/stride. You sign in to Stride in your
browser; there is no key to copy. You choose the project the
agent works in and the projects it can reach, and you can disconnect it at any
time in Stride under Settings → Integrations → Connected apps. The agent
works with your own permissions in each project, and everything it writes is
labeled as your agent's.
Claude Code
claude plugin marketplace add Newlight-Solutions/stride-plugins
claude plugin install stride@stride
Or, inside a session: /plugin marketplace add Newlight-Solutions/stride-plugins,
then /plugin install stride@stride.
Then run /mcp, select plugin:stride:stride, and choose Authenticate.
Your browser opens to sign in to Stride. After that, run /stride:next.
Prefer the server without the commands?
claude mcp add --transport http stride https://www.stride.page/api/mcp/stride
Codex
codex plugin marketplace add Newlight-Solutions/stride-plugins
Then start codex, run /plugins, open Stride AI and install it. Codex
asks you to sign in to Stride when you install. Start a new session, then run
$stride:next.
Cursor
Copy the cursor folder to ~/.cursor/plugins/local/stride and
run Developer: Reload Window. Then open Customize → MCPs, where Stride
shows that it needs to sign in, and follow its prompt; your browser opens to
sign in to Stride. After that, run /stride-next in chat. (On Enterprise,
Cursor loads local plugins only when an admin allows local plugin imports.)
On a Teams or Enterprise plan, an admin can instead import this repository as a team marketplace (Dashboard → Plugins & MCPs → Add Marketplace → Import from Repo), and everyone installs Stride from Customize. Once Stride is listed in the Cursor Marketplace, anyone can install it from Customize: find Stride and select Install.
Prefer the server without the commands? Add it to ~/.cursor/mcp.json:
{
"mcpServers": {
"stride": { "url": "https://www.stride.page/api/mcp/stride" }
}
}
Using an agent key instead
Headless runs (CI, claude -p, codex exec) cannot open a browser to sign in.
For those, create an agent key in Stride under Settings → Integrations and send
it as Authorization: Bearer <key>. Setup for each agent:
Claude Code,
Codex,
Cursor.
This repository
claude-code/,codex/andcursor/are the three packages, each in its agent's own format. They are generated: editsource/and runnpm run build.codex/is also the package submitted to OpenAI's plugin directory (ChatGPT and Codex), as Stride AI, with its review material. It carries its own version, so a directory release leaves the other two packages alone..claude-plugin/marketplace.json,.agents/plugins/marketplace.jsonand.cursor-plugin/marketplace.jsonare the marketplaces the install commands above read.server.jsonis Stride's entry in the MCP Registry.npm run checkrebuilds nothing; it fails if a package has drifted fromsource/or breaks a rule its agent documents. CI runs it on every push.
Questions or problems: [email protected]. Licensed under the MIT License.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.