story-flow

Reviewed at commit ffce5b2 New version, not yet reviewed

Caution: Starts four MCP servers through npx with unpinned versions, so each new package version runs on your machine without review.

Sends data to Context7 (context7.com), Figma (mcp.figma.com), BrowserStack (browserstack.com, only if you configure it), your pull request host (only when you run /prepare-implementation-review)

Agentic development workflow covering planning, analysis, implementation, and QA automation

Made by
Intai
What we checked

Adds slash commands, agents and skills for planning stories, running BDD scenarios in a browser or mobile app and generating Playwright tests. Its hook only watches those runs to show progress in a pane, the status line and toasts. It starts MCP servers for Context7, Chrome DevTools, Playwright, Appium and Figma, and the new /prepare-implementation-review command posts review comments and screenshots to a pull request.

Review details

Static review of the source at the pinned commit; not run. Delta review of the changes since b9558a2; the rest carries over from the review at that commit. The module registers session.start, session.end, prompt.submit, skill.prompt, tool.call (Agent and SubagentHandback results), turn.complete, ui.close and ui.render (Pane), and only observes: it parses story and feature files named in prompts with $.fs.read, keeps run state in atoms and draws a progress pane, status line and toasts; it does not block or change tool calls. plugin.json starts four MCP servers with npx: @upstash/context7-mcp (-y, no version), chrome-devtools-mcp@latest, @playwright/mcp@latest (--no-sandbox, Chrome) and appium-mcp-server@latest, which give Claude control of a browser or device and can reach any site it can. Skills tell Claude to run npx playwright test and, for mobile runs, npx appium server, curl against the BrowserStack WebDriver API and BrowserStackLocal, aws s3 commands against the user's own S3 endpoint and make reseed, using environment variables the user sets; Claude still asks permission for these. The prompt files have no allowed-tools, permissionMode or settings writes. Changes in this version: (1) plugin.json adds a remote HTTP MCP server for Figma at https://mcp.figma.com/mcp, used by the new draft-story command to fetch design context and screenshots for Figma links in a ticket; (2) hook changes are parsing only: scenario ranges like ST-01..ST-05, resuming a finished run, and counting a task result only when the answer leads with Completed or Failed; (3) new command /draft-story and skill draft-story-markdown, which read the ticket through the project's tracker MCP, may drive the app with the Playwright or Appium MCP tools to reproduce a bug, explore the codebase and ask before saving a story file; (4) new command /prepare-implementation-review and skill prepare-implementation-review-comments, which read a pull request, run the changed scenarios and unit specs, take screenshots and axe-core checks of the running app, and then post a review with inline comments and a summary to the pull request through the host's MCP, gh, az, glab or curl with a token from the environment; for GitHub it hosts screenshots by committing them to a pr-evidence orphan branch through a temporary git worktree and pushing it, and it edits and resolves its own earlier comments; --dry-run shows the drafts without posting; (5) /triage-failed-scenarios can now apply, uncommitted, a diff proposed in comments of an earlier issue, skipping ones a human disapproved, so text from issue comments reaches the checkout and depends on who can comment; (6) the execute-bdd-scenario skill adds a @non-parallel tag for recorded specs and sets appium:forceAppLaunch. Posting, pushing and applying go through normal Claude permission prompts. npm packages and the remote MCP servers were not reviewed.

What it uses
  • Reads your prompts: yes
  • Sees tool calls: yes
  • Blocks or changes tool calls: no
  • Reads or writes files: yes
  • Runs programs on your machine: yes
  • Calls the model (uses tokens): no
  • Sends data over the network: yes
  • Stores data: yes
  • Shows something on screen: yes
Hooked events
Made this mod? Add a badge to your README

modsforClaude badge for story-flow

Paste this into your README. The badge says reviewed once the current version has been reviewed, and links readers to this page.

[![Listed on modsforClaude](https://modsforclaude.com/badge/story-flow.svg)](https://modsforclaude.com/mods/story-flow/)

← All mods