AutoPrioritizer AI for Jira Cloud
Last updated: April 22, 2026
AutoPrioritizer AI is built on Atlassian Forge, Atlassian's serverless app platform. The App runs entirely within Atlassian's infrastructure. There are no external servers, databases, or hosting environments owned or operated by us.
api.anthropic.com) for AI analysis.Because the App runs on Forge, we do not manage servers, containers, networking, TLS certificates, or infrastructure patching. Atlassian handles all of this as part of the Forge platform.
The following describes the complete data flow for a prioritization request:
User action (browser)
|
v
Forge Custom UI -> Forge Resolver Function
| |
| |-- Reads issue data from Jira (internal Forge-to-Jira call)
| |-- Reads profile/config from Forge KVS
| |-- Sends ticket fields + criteria to Anthropic API (HTTPS)
| |-- Receives AI recommendation from Anthropic API
| |-- Returns recommendation to frontend for user review
| |
| +-- (On user approval) Writes priority field to Jira issue
v
User sees recommendation, decides whether to apply
Key points:
The App uses Forge's built-in asApp() authentication mechanism to access the Jira REST API. This means:
forge variables set).process.env. It is never sent to the frontend, never included in API responses to the browser, and never logged.api.anthropic.com.The only data persisted by the App is stored in Forge KVS, scoped to the installing Jira instance:
| Data | Storage Key Pattern | Contents |
|---|---|---|
| Prioritization profile | profile-{projectKey} | Criteria weights (numeric), team context (text), priority descriptions (text), custom rules (text), auto-prioritize toggle (boolean) |
| Quick filters | filters-{projectKey} | Array of saved JQL filter snippets and labels |
| Usage statistics | stats-{projectKey} | Total count, monthly usage counts, session history (timestamps and counts), monthly active user lists (account IDs for quota calculation) |
Forge KVS data is encrypted at rest by Atlassian and is tenant-isolated. For details, see Atlassian Forge data storage documentation.
All network communication uses encrypted channels:
| Path | Protocol | Details |
|---|---|---|
| Browser to Forge UI | HTTPS | Standard Atlassian-managed TLS |
| Forge function to Jira API | Internal | Forge-to-Jira calls are internal to Atlassian's infrastructure and do not traverse the public internet |
| Forge function to Anthropic API | HTTPS/TLS | Outbound request to api.anthropic.com over TLS 1.2+ |
No data is transmitted over unencrypted channels.
The App sanitizes all Jira issue data before sending it to the Anthropic API:
These measures reduce the risk of prompt injection, limit data exposure, and keep API payloads to the minimum necessary for analysis.
The App requests the minimum Atlassian scopes required for its functionality:
| Scope | Purpose | How It Is Used |
|---|---|---|
read:jira-work | Read issue fields for AI analysis | Reads summary, description, labels, components, issue type, priority, status, and comments. Used during prioritization requests and issue searches. |
write:jira-work | Update issue priority | Writes only to the priority field on Jira issues. No other fields are modified. All writes require explicit user approval (or admin-enabled auto-prioritize). |
storage:app | Persist App configuration | Stores prioritization profiles, quick filters, and usage statistics in Forge KVS. |
The App is authorized to make outbound HTTP requests only to:
api.anthropic.com -- Anthropic's Claude API for AI analysis.No other external domains are contacted. This is enforced by the Forge platform based on the App's manifest declaration.
| Dependency | Purpose | Data Shared |
|---|---|---|
| Anthropic Claude API | AI-powered ticket analysis | Issue fields (summary, description, labels, components, comments, type, status, current priority) and team prioritization criteria |
Anthropic does not use API inputs or outputs to train its models (per their API Terms of Service).
Because the App runs on Atlassian Forge, it benefits from Forge's built-in security controls:
api.anthropic.com only).If you discover a security vulnerability in AutoPrioritizer AI, please report it responsibly:
Email: support@autoprioritizer.app
Subject line: Security Vulnerability -- AutoPrioritizer AI
Please include:
We will acknowledge receipt within 48 hours and work to address confirmed vulnerabilities promptly. We ask that you not publicly disclose the vulnerability until we have had a reasonable opportunity to address it.
For vulnerabilities in the Atlassian Forge platform itself, please report them through Atlassian's security bug bounty program.
For security-related questions or concerns:
AutoPrioritizer
Email: support@autoprioritizer.app