How to report a security problem in Stintwise, what happens after you do, and how the app is built to limit what a problem could reach. Where something is a limitation rather than a guarantee, this page says so. Reviewing the app for your organisation? For a vendor review gathers the usual questions in one place.
Reporting a vulnerability
Write to security@adigium.com with:
- what you found and where: the dashboard gadget, the dashboard widget, an exported CSV, or this site;
- the steps to reproduce it, on a Jira Cloud site you own or are allowed to test;
- what someone could gain from it: data, access, or a wrong figure.
Mail to that address is delivered by Cloudflare Email Routing to the same Gmail mailbox as support mail. It arrives as ordinary email: there is no encryption key. Please leave out any data that belongs to another organisation.
The same address, with this page as the policy, is in this site's security.txt, the file RFC 9116 describes.
Reporting is written, as all support for this app is. There is no phone line, and no call is needed at any point.
Once Stintwise is listed on the Atlassian Marketplace, you can also report through Atlassian. Its page for reporting a vulnerability in a Marketplace app asks security researchers to "Submit your vulnerability report on any Marketplace App to our public Marketplace bug bounty program at https://bugcrowd.com/atlassianapps", and that programme's own terms apply to a report made there.
What happens next
A report is acknowledged in writing within 1 working day. Working days are those of the support page, and when replies are paused, as the status page says beforehand, so are acknowledgements.
Stintwise works to the fix times Atlassian's Security Bug Fix Policy sets for cloud apps, and they bind it once it is listed on the Atlassian Marketplace:
| Severity | CVSS score | Atlassian's timeframe for resolution |
|---|---|---|
| Critical | Score >= 9.0 | "Must be fixed within 10 days of being reported or triaged." |
| High | Score >= 7.0 | "Must be fixed within 4 weeks of being reported or triaged." |
| Medium | Score >= 4.0 | "Must be fixed within 12 weeks of being reported or triaged." |
| Low | Score < 4.0 | "Must be fixed within 25 weeks of being reported or triaged." |
Each confirmed problem is fixed and released through the same gate as every change: the tests and CI must pass, and the deploy refuses a commit whose CI did not.
Once the fix is deployed, a short note in the changelog says so, and credits you by name if you wish.
Please do not
- test against a Jira site you do not own or are not allowed to test;
- read, change or keep data that is not yours, beyond what you need to show the problem;
- run load or denial-of-service tests. The app runs on Atlassian's infrastructure, so they reach Atlassian and its other customers.
How the app limits what a problem could reach
Each of these is a property of the code, checked by a test on every commit.
- It only reads from Jira. It asks for five permissions:
read:jira-work,read:jira-user,read:board-scope.admin:jira-softwareandread:project:jira, with which it reads, as you and only when you group a chart of a saved filter by board column, the boards that use that filter and their columns, andstorage:appfor its own storage; none lets it write anything in Jira. A test fails the build if awrite:ormanage:permission is added.read:jira-workalone is declared for Atlassian's offline user impersonation, which lets the app read Jira as any person of your site who has access to the app, even while that person is not using it, and Jira then shows it only what that person may see. The app reads that way for every build of a chart, as the person the chart is for, to search and read the work items their own search returns; a test fails the build if another permission is declared for it, or if another part of the code reads that way. The gadget's settings are saved by Jira with the gadget, when someone who can edit the dashboard saves them. - It runs no servers of its own. Its code runs on Atlassian Forge, and it keeps its charts in Forge's storage; the view and options you last chose on a tile are kept by your browser (privacy). Its one outside host is PostHog's analytics endpoint in the European Union, which your administrator can switch off (how).
- It stores nothing that names a person, and of a work item's words only its labels, its components' names and its team's name, which split the chart and which the CSV export writes: no summaries, descriptions, comments, attachments, custom field values (but a sprint's id, board, state and dates and the name of the team in Jira's Team field), sprint names and goals, assignees, reporters, change authors or account ids. A chart built from one person's search is shown to another only when Jira, asked as that other person, confirms their own searches return exactly its work items and that they can open every one, and never to or from someone who can edit its saved filter or its dashboard, nor for a filter that names another filter or uses an app's function; which chart a person was shown is remembered by a key made from the site and the account with a secret the app keeps, which is not the account id. Every field it stores is on a closed list that a test checks (privacy).
- A chart is served only to someone who can open what it lists, checked as that person, and, when it was built from someone else's search, whose own searches Jira confirms return exactly its work items, with the limits stated under who sees what. This is tested, and not yet checked on a real site with two accounts of different access.
- Settings sent from a dashboard are not trusted. A widget's settings arrive with each request and are checked on each one: a saved filter the viewer cannot open is refused, and a status name is quoted as one string in the query, so it cannot change what the query asks.
- An exported CSV cannot run a formula in your spreadsheet. A cell a spreadsheet would run is written with a leading apostrophe, the spreadsheets' own mark for text.
- It holds no Atlassian password or token, and reaches Jira only through Forge.
- Its production dependencies are checked. On every push, and once a week, CI fails if a production dependency has a known high or critical advisory, and the app's deploy refuses a commit whose CI failed. An advisory passes only if it is listed as an exception with its reason and an end date, and then only until that date and while the tile's built code does not contain the affected package's name. Moderate and low advisories, development tools, and advisories not yet published are not covered.
What is not in place
- Stintwise is not yet in Atlassian's Marketplace bug bounty programme, and has had no third-party penetration test.
- The developer holds no SOC 2 or ISO 27001 certification. Atlassian states that the Forge platform the app runs on "is SOC 2 attested and ISO 27001 certified"; that is Atlassian's, not the developer's.
- Cloud Fortified, the badge some apps in this category show, cannot be obtained: Atlassian's page says "We are no longer accepting new submissions for the Cloud Fortified Apps Program. Cloud Fortified will be phased out by the end of CY2026."
- One alert rule is all the monitoring there is. Atlassian's developer console emails the developer when, over one hour, the share of the app's invocations that succeed falls below 95%, as major, or below 80%, as critical. It watches nothing else: not the figures the tile states, so a chart that loads with a wrong number raises no alert. Nor does it see a chart that fails to build or a tile that cannot reach Jira: the tile says so in its own message, and Atlassian counts as failures only errors the app does not catch, timeouts and running out of memory. The alert is an email, and a person then reads it. A problem in Stintwise itself that affects everyone is posted on the status page once the developer knows of it.
- What the app writes to Atlassian's logs, and how your administrator stops sharing them, is on the privacy page under Logs.
This site
Every page of this site, the try page among them, is sent a Content-Security-Policy that lets it load only files from this site and open no connection.
For a vendor review
The questions a security or privacy review of an app usually asks, each in a few sentences, with the page that gives the detail. What the developer has not yet decided is listed last, and no page of this site states it until it is decided.
Who makes it
One person builds and maintains Stintwise. Its support is written, by email, as the support page describes.
Terms and data processing
Stintwise is not yet listed on the Atlassian Marketplace, so no licence agreement is in force. Once it is listed it will use Atlassian's standard End User Agreement, unmodified (terms). No data processing agreement from the developer is published yet. The app runs on Atlassian Forge, whose Data Processing Addendum is, in Atlassian's words, a contract between "Atlassian and Forge developers": it "is incorporated by reference in the Forge Terms, which all Forge developers must comply with", and "also incorporates the EU standard contractual clauses (EU SCCs)" (Atlassian, read 10 October 2026). It is not an agreement between you and the developer.
Where data is kept
Everything the app keeps in its own storage is in Atlassian Forge's hosted storage, in the region Atlassian allocates to your site. The app operates no servers, database or infrastructure of its own (where data lives). Atlassian says that data in that storage "is held in the same location as the host Atlassian app, and moves with it when an admin migrates that app to another location" (Atlassian, read 10 October 2026), and that Forge "will aim to execute its invocations from the same location as the host Atlassian app" but "may sometimes execute an app's invocation from a location other than the location where the host Atlassian app is" (Atlassian, read 10 October 2026). For data residency, Stintwise counts as in scope everything it keeps in that storage, as the privacy page's What the app stores lists it; its four product events and its log lines are not, and its manifest declares its one outside host, PostHog's analytics endpoint in the European Union, as receiving no in-scope data. Three things on that list are not in the app's storage: the gadget's settings, which Jira keeps with your dashboard, a Stintwise action's settings, which Jira Automation keeps with its rule, and the choices each person makes on a tile, which their browser keeps.
Sub-processors
Atlassian, for hosting, storage and billing, and PostHog, for product analytics in the European Union, whose events carry a salted hash of your site's cloud id and nothing that identifies a person or a work item. There are no others for the app (sub-processors). Mail you send to stintwise@adigium.com or security@adigium.com is delivered by Cloudflare Email Routing to a Gmail mailbox, so Cloudflare and Google handle what you write there.
Personal data
The app stores no account id and no name or email address of any person (what it never stores). Jira sends some with what the app asks for, unasked: who made each status change, and anything the tool that made it attached; when Jira will not filter a work item's history, the whole of it, which can name the people it was assigned to; the saved filter's owner and sharing; the dashboard's owner and sharing; a board's location, which can name the person whose profile holds it, and the names of the boards that use a saved filter; and, while a new gadget's settings are open, the configuring person's profile. The app discards it in memory as the response is read, keeping only that person's time zone, which Jira saves with the gadget's settings. A chart built from one person's search is shown to another only when Jira, asked as that other person, confirms their own searches return exactly its work items and that they can open every one, and never to or from someone who can edit its saved filter or its dashboard, nor for a filter that names another filter or uses an app's function; which chart a person was shown is remembered by a key made from the site and the account with a secret the app keeps; it is not the account id.
How long data is kept
Everything the app writes for a chart but what its next build starts from expires 25 hours after it is written, and, since Atlassian's storage may take up to 48 hours to remove an expired record, the app can no longer read it 73 hours after a chart was built, at the latest. What a chart's next build starts from, every work item its searches returned with what was read of it, expires 97 hours after it is written, so that the next morning's first view, or the first after a weekend, can start from the last chart built before it, and the app can no longer read it 145 hours after it is written, at the latest. One random secret per installation, and the random key that seals the JQL queries typed into the settings, are kept while the app is installed. What Atlassian keeps after that, and after the app is uninstalled, is Atlassian's, quoted on the privacy page. PostHog keeps the product events available for a year on the free plan this project uses. Of the app's log lines, Atlassian says "Logs are only available for the past 30 days" (logs).
Who can see what
The app asks for five permissions: read:jira-work, read:jira-user, read:board-scope.admin:jira-software and
read:project:jira, with which it reads, as the person viewing and only when they group a chart of a saved filter
by board column, the boards that use that filter and their columns, and storage:app for its own storage; none
lets it write anything in Jira. Of them, read:jira-work alone is declared for Atlassian's
offline user impersonation, with which "the app can impersonate any user within the context where it is
installed" (Atlassian, read 10 October 2026):
the app may read Jira as any person of your site who has access to it, even while that person is not using
it, and is shown only what Jira lets that person see. It reads that way for every build of a chart, as the
person the chart is for, to search and read the work items their own search returns. A chart is built from
what one person's saved filter, project or JQL query returns for them, searched as them; it is served to
another person only after the app confirms, as that person, that their own searches return exactly its work
items, and before a chart, a drill panel or an export is served the app
checks, as that person, that they can open each saved filter, browse each project, and have Jira's strict
check accept each JQL query from them, with no error and no warning, and that they can open every work
item the chart lists, with the limits that who can see a chart states:
Jira's check of a query does not look at the people it names, its labels, text searches or function
arguments, so those words are shown to everyone it runs for. The
tile names a saved filter only to someone Jira lets open it; the settings Jira hands everyone who can
see a dashboard hold each saved filter's id but no saved filter's name once saved after the update of
8 October 2026, a project by its id, and a JQL query sealed with a key only the app holds, so they show no
query's text; they also hold the days the whole team is off as listed, which the app cannot tell from one
person's leave, and each source's holiday list, also for a source a person cannot open (what they hold). An export names a work
item's parent only where that person's own search returns it. A work item's team is named as Jira's Team
field gives it to the app; whether Jira would show that person the team is not checked. A summary that a Stintwise
action in a Jira Automation rule writes is counted as the person whose connection the action uses, and the rule sends it
to whoever it names, outside Jira's permissions
(a summary a rule sends). Apart from the product events and
the log lines, nothing the app computes reaches its author. This is tested, and not yet checked on a
real site with two accounts of different access.
Logs
The app writes a line to Atlassian's Forge logs when a chart's build is queued, starts and finishes, when a build read some history whole, when a build reads the site's status list again, when a chart is served after asking Jira whether you can still open its work items, when a chart built from someone else's search is checked against your own searches, when a chart's build reads a search in full where it could have started from an earlier build, when Jira asks a build to slow down, when Jira's search returns work items out of the order a build asked for, when Jira says the app's hourly limit is used up or that little of that limit is left, when the app holds a new chart back so as not to use up Jira's hourly limit, when Jira would not let the app act as you for a chart's build, when a saved filter was edited between your view and its chart's build, when something fails, and, in the app's development environment only, when it prices a kind of request: an event name or an error code and, depending on the event, counts, durations, times and codes, how much of the app's hourly limit Jira says is left and how much its requests used, and for a fault in the app's own code its message and the first lines of its stack, but never the message of a failure Jira reports (every field). Atlassian shares these lines with the app's author unless your administrator turns off "Logs access".
Vulnerabilities
Report one to security@adigium.com, as Reporting a vulnerability above says; What happens next gives the acknowledgement and Atlassian's fix times. On every push, and once a week, CI fails if a production dependency has a known high or critical advisory not listed as an exception with its reason and an end date, and the deploy refuses a commit whose CI failed; How the app limits what a problem could reach above says what this does not cover.
Incidents
An alert rule watches the production app; the status page says what it sees and what it does not. A problem in Stintwise itself that affects everyone is posted on the status page once the developer knows of it. Atlassian's guidelines for a security incident in a Marketplace app, which Stintwise is not yet, say: "When possible, Atlassian recommends notifying customers within 72 hours of identification of an incident" (Atlassian, read 10 October 2026). The developer has not yet stated a notice time of its own.
Certifications and independent checks
The developer holds no SOC 2 or ISO 27001 certification. Atlassian states that the Forge platform the app runs on "is SOC 2 attested and ISO 27001 certified"; that is Atlassian's, not the developer's. Stintwise has had no third-party penetration test, and is not yet in Atlassian's Marketplace bug bounty programme. Two accounts of different access have not yet been checked on a real site (status).
Not yet decided
These are the developer's to decide or confirm, and no page of this site states them until they are: the name, country and address of the party you contract with; a data processing agreement between that party and you; whether one is signed with PostHog; insurance; how soon you are told of a security incident that affects your data; and how the developer's own accounts and computers are secured.