Frequently Changed Logs



  • Adds a new command danger local.

This command will look between the current branch and master and use that to evaluate a dangerfile. This is aimed specifically at tools like git commit hooks, and for people who don’t do code review.

danger.github will be falsy in this context, so you could share a dangerfile between danger local and danger ci.

When I thought about how to use it on Danger JS, I opted to make another Dangerfile and import it at the end of the main Dangerfile. This new Dangerfile only contains rules which can run with just danger.git, e.g. CHANGELOG/README checks. I called it dangerfile.lite.ts.

Our setup looks like:

  "scripts": {
    "prepush": "yarn build; yarn danger:prepush",
    "danger:prepush": "yarn danger local --dangerfile dangerfile.lite.ts"
    // [...]

You’ll need to have husky installed for this to work. - @orta

  • STDOUT formatting has been improved, which is the terminal only version of Danger’s typical GitHub comment style system. It’s used in danger pr, danger ci --stdout and danger local. - @orta
  • Exposed a get file contents for the platform abstraction so that Peril can work on many platforms in the future - @orta


  • Added support for Bitrise as a CI Provider - @tychota



  • Paginate for issues - @orta


  • Added support for as a CI Provider - @fbartho


  • Don’t log ENV vars during a run - thanks @samdmarshall. - @orta


  • Bug fixes and debug improvements. If you’re interested run danger with DEBUG="*" yarn danger [etc] and you’ll get a lot of output. This should make it much easier to understand what’s going on. - @orta


  • Updates to the CLI user interface. Breaking changes for everyone.

TLDR - change yarn danger to yarn danger ci.

Danger JS has been fighting an uphill battle for a while with respects to CLI naming, and duplication of work. So, now it’s been simplified. There are four user facing commands:

  • danger init - Helps you get started with Danger
  • danger ci - Runs Danger on CI
  • danger process - Like ci but lets another process handle evaluating a Dangerfile
  • danger pr - Runs your local Dangerfile against an existing GitHub PR. Will not post on the PR

This release deprecates running danger on it’s own, so if you have yarn danger then move that be yarn danger ci.

Each command name is now much more obvious in it intentions, I’ve heard many times that people aren’t sure what commands do and it’s is still even worse in Danger ruby. I figure now is as good a time as any a good time to call it a clean slate.

On a positive note, I gave all of the help screens an update and tried to improve language where I could.


  • Fix to danger pr and danger infinite looping - @orta


  • Add a note in danger pr if you don’t have a token set up - @orta
  • Bunch of docs updates - @orta


  • Fix Codeship integration - @caffodian
  • Updates documentation dependencies - @orta
  • Fixes to running danger with params - @orta
  • Fixes for danger pr not acting like danger WRT async code - @orta
  • Fixes tsconfig.json parse to be JSON5 friendly - [@gantman][]
  • Fixes for danger.github.thisPR to use the base metadata for a PR, I’m too used to branch workflows - @orta


  • Updates dependencies - @orta
  • Link to the build URL if Danger can find it in the CI env - @orta
  • Removes the “couldn’t post a status” message - @orta


  • The TS compiler will force a module type of commonjs when transpiling the Dangerfile - @orta


  • Adds a CLI option for a unique Danger ID per run to danger and danger process, so you can have multiple Danger comments on the same PR. - @orta

#2.1.1 - 2.1.2 - 2.1.3

  • Fixes/Improvements for danger init - @orta


  • Adds a new command for getting set up: danger init - @orta
  • Fix double negative in documentation. @dfalling
  • Fix gloabally typo in documentation. @dfalling

#2.0.2 - 2.0.3

  • Adds a warning when you try to import Danger when you’re not in a Dangerfile - @orta
  • Exports the current process env to the danger run subprocess - @orta


  • Potential fixes for CLI sub-commands not running when packaging danger - @orta


  • Fixes the danger.js.flow fix to handle exports correctly, you probably need to add .*/node_modules/danger/distribution/danger.js.flow to the [libs] section of your settings for it to work though - @orta


  • Fixes a bug with danger.github.utils in that it didn’t work as of b1, and now it does :+1: - @orta
  • Ships a danger.js.flow in the root of the project, this may be enough to support flow typing, thanks to @joarwilk and flowgen - @orta


  • Converts the command danger (and danger run) to use danger process under the hood. What does this do?

    • Dangerfile evaluation is in a separate process, run without a vm sandbox. This fixes the async problem which we created schedule for. Previously, any async work in your Dangerfile needed to be declared to Danger so that it knew when all of the work had finished. Now that the running happens inside another process, we can use the on_exit calls of the process to know that all work is done. So, in Danger (not in Peril) async code will work just like inside a traditional node app.
    • Makes danger process a first class citizen. This is awesome because there will be reliable support for other languages like danger-swift, danger-go and more to come.
    • The danger process system is now codified in types, so it’s really easy to document on the website.
  • Adds a --js and --json option to danger pr which shows the output in a way that works with danger process. This means you can preview the data for any pull request.



Moves away from vm2 to a require-based Dangerfile runner. This removes the sandboxing aspect of the Dangerfile completely, but the sandboxing was mainly for Peril and I have a plan for that.

I would like to move the main parts of Danger JS to also work like danger process, so I’ll be continuing to work as a alpha for a bit more. One interesting side-effect of this could be that I can remove schedule from the DSL. I’ve not tested it yet though. Turns out this change is real hard to write tests for. I’ve made #394 for that.


#2.0.0-alpha.18 - 19

  • Moves internal methods away from Sync to avoid problems when running in Peril - @ashfurrow
  • Passes through non-zero exit codes from danger process runs - @ashfurrow


  • Improve CircleCI PR detection


Some UX fixes:

  • Don’t show warnings about not setting a commit status (unless in verbose) - @orta
  • Delete duplicate Danger message, due to fast Peril edits - @orta
  • Show Peril in the commit status if inside Peril, not just Danger - @orta
  • [internal] Tightened the typings on the commands, and abstracted them to share some code - @orta


  • Updates diffForFile, JSONPatchForFile, and JSONDiffForFile to include created and removed files - #368 - bdotdub


  • Adds a blank project generated in travis 8 to test no-babel or TS integration - @orta
  • Improvements to danger process logging, and build fails correctly #363 - @orta


  • Improve the error handling around the babel API - #357 - @orta
  • Move back to the original URLs for diffs, instead of relying on PR metadata - @orta
  • Updates the types for schedule to be more accepting of what it actually takes - @orta


  • Fixed #348 invalid json response body error on generating a diff - felipesabino
  • Potential fix for ^ that works with Peril also - @orta


  • Doh, makes the danger process command actually available via the CLI - @orta


  • Adds a danger process command, this command takes amn argument of a process to run which expects the Danger DSL as JSON in STDIN, and will post a DangerResults object to it’s STDOUT. This frees up another process to do whatever they want. So, others can make their own Danger runner.

An example of this is Danger Swift. It takes a JSON document via STDIN, compiles and evaluates a Swift file then passes the results back to danger process via STDOUT.

Another example is this simple Ruby script:

    #!/usr/bin/env ruby

  require 'json'
  dsl_json = STDIN.tty? ? 'Cannot read from STDIN' : $
  danger = JSON.parse(dsl_json)
  results = { warnings: [], messages:[], fails: [], markdowns: [] }

  if "Hello world"
    results.messages << { message: "Hey there" }

  require 'json'

Which is basically Ruby Danger in ~10LOC. Lols.

This is the first release of the command, it’s pretty untested, but it does work. - @orta


  • Uses the Babel 7 alpha for all source compilation with JS, Flow+JS and TS. This worked without any changes to our internal infra which is pretty awesome. All TS tests passed. Babel 7 is still in alpha, but so is Danger 2.0 - so I’m happy to keep Danger in a pretty long alpha, till at least Babel 7 is in beta.

It also still supports using TypeScript via the “typescript" module, if you have that installed. - @orta

  • danger.github.thisPR now uses the PR’s head, not base - @orta


  • Uses the GitHub diff_url instead of the diff version header, as it conflicted with Peril - @orta
  • Handle exceptions in Dangerfile and report them as failures in Danger results - macklinu


  • Expose a Promise object to the external GitHub API - @orta


  • Allow running a dangerfile entirely from memory using the Executor API - @orta


  • Removes the jest-* dependencies - @orta


  • Support a vm2 based Dangerfile runner as an alternative to the jest infrastructure. There are a few main reasons for this:

    • I haven’t been able to completely understand how Jest’s internals work around all of the code-eval and pre-requisite setup, which has made it hard to work on some more complex Peril features.
    • Jest releases are every few months, personally I love this as a user of Jest, as an API consumer it can be difficult to get changes shipped.
    • The fact that both Danger/Jest make runtime changes means that you need to update them together
    • I have commit access to vm2, so getting changes done is easy

I like to think of it as having gone from Jest’s runner which is a massive toolbox, to vm2 which is a tiny toolbox where I’ll have to add a bunch of new tools to get everything working.

The massive downside to this is that Danger now has to have support for transpiling via Babel, or from TypeScript unlike before, where it was a freebie inside Jest. Jest handled this so well. This means that a Dangerfile which used to "just work” with no config may not. Thus, IMO, this is a breaking major semver.

Is it likely that you need to make any changes? So far, it seems probably not. At least all of the tests with Dangerfiles original from the older Jest runner pass with the new version.

This is an alpha release, because it’s knowingly shipped with some breakages around babel support, specifically:

  • Babel parsing of relative imports in a Dangerfile aren’t working
  • Some of the features which require the regeneratorRuntime to be set up aren’t working yet

Those are blockers on a 2.0.0 release.


  • Exposes an internal API for reading a file from a GitHub repo as danger.github.utils.fileContents - @orta

Ideally this is what you should be using in plugins to read files, it’s what Danger uses throughout the codebase internally. This means that your plugin/dangerfile doesn’t need to rely on running on the CI when instead it could run via the GitHub API.

  • Update prettier - @orta
  • Removes dtslint as a dependency - sapegin/orta


  • Support retrive paginated pull request commit list - kwonoj
  • Add support for VSTS CI - mlabrum
  • Remove the DSL duplication on the danger export, it wasn’t needed or used. - @orta
  • Update to TypeScript 2.4.x - @orta
  • Rename github test static_file to remove : from the filename to fix a checkout issue on windows - mlabrum


Hello readers! This represents a general stability for Danger. It is mainly a documentation release, as it corresponds to being generally available. I made the initial commit back in 20 Aug 2016 and now it’s 30th June 2017. It’s awesome to look back through the CHANGELOG and see how things have changed.

You can find out a lot more about the 1.0, and Danger’s history on my Artsy blog post on the Danger 1.0.

  • Adds inline docs for all CI providers - @orta


  • Use HTTP for the GitHub status check target URL - macklinu
  • Correct some examples in node-app - clintam
  • Add support for buddybuild CI - benkraus/clintam
  • Add support for GithHub Apps API (no GET /user) - clintam


  • Posts status reports for passing/failing builds, if the account for danger has access - @orta
  • Adds prettier to the codebase - @orta
  • Converts a bunch of Danger’s dangerfile into a plugin - danger-plugin-yarn - @orta

This is my first non-trivial plugin, based on infrastructure from @macklinu. Plugins are looking great, you can get some info at

  • Docs updates for the website - @orta


  • Fix danger pr commands are not running on windows - kwonoj
  • Fix broken link in getting started docs - frozegnome
  • Do not delete comment written from user have same userid for danger - kwonoj
  • Fix link to jest in getting started docs - palleas
  • Fix yarn install instruction in getting started docs - palleas


  • Update to Jest 20 - macklinu
  • Change the danger.d.ts to use module exports instead of globals - @orta
  • Render markdown inside message(), warn(), and fail() messages. - macklinu

An example:

fail(`Missing Test Files:

- \`src/lib/components/artist/artworks/__tests__/index-tests.tsx\`
- \`src/lib/components/artwork_grids/__tests__/infinite_scroll_grid-tests.tsx\`
- \`src/lib/containers/__tests__/works_for_you-tests.tsx\`

If these files are supposed to not exist, please update your PR body to include "Skip New Tests".`)

Will result in:

:no_entry_sign: Missing Test Files: * `src/lib/components/artist/artworks/__tests__/index-tests.tsx` * `src/lib/components/artwork_grids/__tests__/infinite_scroll_grid-tests.tsx` * `src/lib/containers/__tests__/works_for_you-tests.tsx` If these files are supposed to not exist, please update your PR body to include “Skip New Tests”.


  • Adds github.api. This is a fully authenticated client from the github npm module. - @orta

An easy example of it’s usage would be using Danger to add a label to your PR. Note that Danger will have the permissions for your account, so for OSS repos - this won’t work.

    owner: "danger",
    repo: "danger-js",
    labels: ["Danger Passed"],

Yeah, that’s a bit verbose, I agree. So, there’s also github.thisPR which should simplify that. It aims to provide a lot of the values for the current PR to use with the API.

  const github = danger.github
  github.api.issues.addLabels({ ...github.thisPR, labels: ["Danger Passed"] })

You could use this API for a bunch of things, here’s some quick ideas:

  • Request specific reviewers when specific files change (api.pullRequests.createReviewRequest)
  • Add a label for when something passes or fails (api.issues.addLabels)
  • Verifying if someone is in your org? (api.orgs.checkMembership)
  • Updating Project tickets to show they have a PR (api.projects.updateProject)


  • Enhancements to danger.git.diffForFile() - @namuol

    • Removed diffTypes second argument in favor of result.added and result.removed
    • Added result.before and result.after for easy access to full contents of the original & updated file
    • danger.git.diffForFile is now an async function

#### TL;DR:

  // In danger 0.16.0:
  const fullDiff = danger.git.diffForFile('foo.js')
  const addedLines = danger.git.diffForFile('foo.js', ['add'])
  const removedLines = danger.git.diffForFile('foo.js', ['del'])

  // In the latest version:
  const diff = await danger.git.diffForFile('foo.js')
  const fullDiff = diff.diff
  const addedLines = diff.added
  const removedLines = diff.removed
  const beforeFileContents = diff.before
  const afterFileContents = diff.after
  • Update internal test fixture generation docs - namuol


  • Adds a diffTypes option to diffForFile - alex3165
  • Add Buildkite CI source - jacobwgillespie


  • When a Dangerfile fails to eval, send a message to the PR - @orta


  • Updated jest-* dependencies to 19.x - @orta

Updating the jest-* dependencies seems to be exhibiting strange behavior in tests for windows if you update, and use windows, can you please confirm that everything is 👍

  • Added type shapings to JSONPatchForFile - @orta
  • Replaced deprecated lodash.isarray package with Array.isArray - damassi


  • Moved @types/chalk from dependencies to devDependencies - @orta
  • Killed some stray console logs - @orta
  • Updated the danger.d.ts - @orta


  • TypeScript Dangerfiles are now support in Danger - @orta

We use TypeScript in Danger, and a lot of my work in Artsy now uses TypeScript (see: JS2017 at Artsy), so I wanted to explore using TypeScript in Dangerfiles.

This is built on top of Jest’s custom transformers, so if you are already using Jest with TypeScript, then you can change the dangerfile.js to dangerfile.ts and nothing should need changing ( except that you might have new warnings/errors ) (note: in changing this for Danger, I had to also add the dangerfile.ts to the "exclude" section of the tsconfig.json so that it didn’t change the project’s root folder.)

This repo is now using both a babel Dangerfile (running on Circle CI) and a TypeScript one (running on Travis) to ensure that we don’t accidentally break either.

  • Created a new danger.d.ts for VS Code users to get auto-completion etc - @orta
  • Added a two new git DSL functions: git.JSONDiffForFile(filename) and git.JSONPatchForFile(filename).

    • git.JSONPatchForFile

    This will generate a rfc6902 JSON patch between two files inside your repo. These patch files are useful as a standard, but are pretty tricky to work with in something like a Dangerfile, where rule terseness takes priority. * git.JSONDiffForFile

    This uses JSONPatchForFile to generate an object that represents all changes inside a Dangerfile as a single object, with keys for the changed paths. For example with a change like this:

      "dependencies": {
        "babel-polyfill": "^6.20.0",
     +  "chalk": "^1.1.1",
        "commander": "^2.9.0",
        "debug": "^2.6.0"

    You could become aware of what has changed with a Dangerfile in a schedule’d function like:

    const packageDiff = await git.JSONDiffForFile("package.json")
    if (packageDiff.dependencies) {
      const deps = packageDiff.dependencies
      deps.added   // ["chalk"],
      deps.removed // []
      deps.after   // { "babel-polyfill": "^6.20.0", "chalk": "^1.1.1", "commander": "^2.9.0", "debug": "^2.6.0" }
      deps.before  // { "babel-polyfill": "^6.20.0", "commander": "^2.9.0", "debug": "^2.6.0" }

    The keys: added and removed only exist on the object if: * before and after are both objects - in which case added and removed are the added or removed keys * before and after are both arrays - in which case added and removed are the added or removed values

  • Exposed all global functions ( like warn, fail, git, schedule, … ) on the danger object. - @orta

This is specifically to simplify building library code. It should not affect end-users. If you want to look at making a Danger JS Plugin, I’d recommend exposing a function which takes the danger object and working from that. If you’re interested, there is an active discussion on plugin support in the DangerJS issues.

  • Improves messaging to the terminal - @orta
  • Adds the ability to not have Danger post to GitHub via a flag: danger run --text-only - @orta
  • Fix a crasher with prs.find #181 - @orta


  • Add danger.utils DSL, which includes danger.utils.href() and danger.utils.sentence() - macklinu

We were finding that a lot of Dangerfiles needed similar functions, so we’ve added a utils object to offer functions that are going to be used across the board. If you can think of more functions you use, we’d love to add them. Ideally you shouldn’t need to use anything but Danger + utils to write your Dangerfiles.

  danger.utils.href("", "Danger") // <a href="">Danger</a>
  danger.utils.sentence(["A", "B", "C"]) // "A, B and C"
  • Adds danger.github.utils - which currently has only one function: fileLinks - @orta

Most of the time people are working with a list of files (e.g. modified, or created) and then want to present clickable links to those. As the logic to figure the URLs is very GitHub specific, we’ve moved that into it’s own object with space to grow.

  const files = danger.git.modified_files // ["lib/component/a.ts", "lib/component/b.ts"]
  const links = danger.github.utils.fileLinks(files) // "<a href='...'>a</a> and <a href='...'>b</a>"
  warn(`These files have changes: ${links}`)


  • Add support for Drone - gabro


  • Added support for handling async code in a Dangerfile - deecewan

This is still a bit of a work in progress, however, there is a new function added to the DSL: schedule.

A Dangerfile is evaluated as a script, and so async code has not worked out of the box. With the schedule function you can now register a section of code to evaluate across multiple tick cycles.

schedule currently handles two types of arguments, either a promise or a function with a resolve arg. Assuming you have a working Babel setup for this inside your project, you can run a Dangerfile like this:

  schedule(async () => {
    const thing = await asyncAction()
    if (thing) {
      warn("After Async Function")

Or if you wanted something simpler,

  schedule(resolved => {
    if (failed) {
      fail("Failed to run")
  • Adds new GitHub DSL elements - deecewan

  • danger.github.issue - As a PR is an issue in GitHub terminology, the issue contains a bit more metadata. Mainly labels, so if you want to know what labels are applied to a PR, use danger.github.issue.labels

  • - Find out about your reviews in the new GitHub Reviewer systems,

  • danger.github.requested_reviewers - Find out who has been requested to review a PR.

  • Updated TypeScript and Jest dependencies - @orta

  • Add support for Github Enterprise via DANGER_GITHUB_API_BASE_URL env var - mashbourne

#0.11.3 - 0.11.5

  • Internal changes for usage with Peril - @orta

  • Add danger pr --repl, which drops into a Node.js REPL after evaluating the dangerfile - macklinu

  • Add support for Codeship - deecewan

#0.11.0 - 0.11.2


  • Builds which only use markdown now only show the markdown, and no violations table is shown - mxstbr


  • Adds support for running Danger against a PR locally - @orta

The workflow is that you find a PR that exhibits the behavior you’d like Danger to run against, then edit the local Dangerfile.js and run yarn run danger pr

This will post the results to your console, instead of on the PR itself.

  • Danger changes to your Dangerfile are not persisted after the run - @orta
  • Add summary comment for danger message - kwonoj
  • Add jest-environment-node to the Package.json - @orta


  • Adds support for git.commits and github.commits - @orta

Why two? Well github.commits contains a bunch of github specific metadata ( e.g. GitHub user creds, commit comment counts. ) Chances are, you’re always going to use git.commits however if you want more rich data, the GitHub one is available too. Here’s an example:

const merges = git.commits.filter(commit => commit.message.include("Merge Master"))
if (merges.length) {
  fail("Please rebase your PR")
  • Support custom dangerfile via -d commandline arg - kwonoj
  • Allow debug dump output via DEBUG=danger:* environment variable - kwonoj
  • Adds surf-build ci provider - kwonoj
  • Forward environment variables to external module constructor - kwonoj


  • Support danger run -ci to specify external CI provider - kwonoj
  • Adds --verbose to danger, which for now will echo out all the URLs Danger has requested - @orta
  • Migrate codebase into TypeScript from flow - kwonoj
  • Handle removing all sorts of import types for Danger in the Dangerfile - @orta


  • A failing network request will raise an error - @orta
  • Fix Dangerfile parsing which broke due to Peril related changes - @orta
  • Tweak the npmignore, ship less random stuff to others - @orta


  • Fixes to the shipped Flow/TS definitions - @orta
  • Adds more functions the the internal Danger GitHub client - @orta
  • Infrastructure work to allow Peril to run a Dangerfile - @orta
  • Upgrade outdated ESLint packages - macklinu
  • Enhance Windows OS compatibility - kwonoj


  • Set exit code to 1 when running danger throws an error - macklinu
  • Add Jenkins CI source - macklinu
  • Add .editorconfig - macklinu
  • Adds jest-runtime to the dependencies - @orta


  • You can build and run in vscode using your own custom env/development.env file. This is useful because you can use the debugger against a real PR. See env/development.env.example for syntax. - @orta

  • Uses jest-transform and jest-runtime to eval and apply babel transforms.

This does two things, makes it feasible to do hosted-danger and makes it possible to write your Dangerfile in a way that’s consistent with the rest of your JavaScript. - @orta

  • Add tests directory to .npmignore - macklinu
  • Update to Jest 18 - macklinu


  • Brings back the ability to emulate a fake CI run locally via danger - @orta


  • Makes babel-polyfill a direct dependency, this is because it is actually an implicit dependency in the app. I’m not sure how I feel about this, I guess if we use a part of it in the babel translation of a user’s Dangerfile them I’m OK with it. - @orta

#0.6.6 - 0.6.7 - 0.6.8

  • Ship flow annotations with the npm module - @orta


  • Adds more node instances to travis - romanki + orta
  • Adds support for Semaphore CI - @orta


  • The env vars DANGER_TEST_REPO and DANGER_TEST_PR will allow you initialize the FakeCI with a repo of your choice. See for more info
  • Improved error messaging around not including a DANGER_GITHUB_API_TOKEN in the ENV - nsfmc / orta
  • Adds support for getting the diff for a specific file from git: e.g.
const changelogDiff = danger.git.diffForFile("")
const contributorName =
if (changelogDiff && changelogDiff.indexOf(contributorName) === -1) {
  warn("Please add your GitHub name to the changelog entry, so we can attribute you.")


  • Does not break commonmark on GitHub - @orta
  • upgrades to flow 0.35.0 and fixes associated type errors in covariant/invariant interfaces - nsfmc
  • omits flow requirement for new test files - nsfmc
  • adds support for circleci - nsfmc
  • defines CISource properties in flow as read-only - nsfmc


  • ->, I’ve also created interfaces for them - @orta
  • warn, message, markdown are all ported over to DangerJS - @orta
  • Shows a HTML table for Danger message - @orta
  • Now offers a Flow-typed definition file, it’s not shipped to their repo yet, you can make it by npm run export-flowtype - @orta
  • Started turning this into a real project by adding tests - @orta


  • Changes some files casing, added some logs, a bit of error reporting, and verifying everything works through npm - @orta


  • Danger edit an existing post, and delete it when it’s not relevant - @orta


  • Danger will post a comment on a GitHub PR with any Fails - @orta


OK, first usable for others version. Only supports GitHub and Travis CI.

You can run by doing:


Make sure you set a DANGER_GITHUB_API_TOKEN on your CI - see the Ruby guide for that.

Then you can make a dangerfile.js (has to be lowercase, deal with it.) It has access to a whopping 2 DSL attributes.

fail(message: string)

pr probably won’t be sticking around for the long run, but if you’re using a 0.0.2 release, you should be OK with that. It’s the full metadata of the PR, so this JSON file. git currently has:


which are string arrays of files.

fail(message: string) will let you raise an error, and will make the process return 1 after the parsing has finished.

Overall: your Dangerfile should look something like:

import { danger } from "danger"

const hasChangelog = danger.git.modified_files.includes("")
if (!hasChangelog) {
  fail("No Changelog changes!")

That should do ya. I think. This doesn’t support babel, and I haven’t explored using other modules etc, so…



Not usable for others, only stubs of classes etc. - @orta

Got improvements? Help improve this document via sending PRs.