ContextGuard AI

An intelligent release risk engine for Jira that analyzes dynamic issue signals, dependency context, and posts automated ADF audit comments to prevent deployment failures.

Comment

Inspiration

In fast-paced engineering teams, critical hotfixes and deployments often slip through release pipelines without full visibility into missing architectural documentation, unlinked dependencies, or high-risk scope changes. ContextGuard AI was built to bring real-time risk intelligence directly inside Jira issue workflows.

What it does

ContextGuard AI embeds an interactive UI Kit panel into Jira issues that automatically evaluates deployment risk vectors before code reaches production:

  • Dynamic Risk Scoring: Queries live Atlassian REST APIs to analyze issue priority, high-risk summary keywords, and dependency traceability to generate a 0–100 risk score.
  • Visual Vector Breakdown: Highlights specific risk signals (e.g., Critical priority, unlinked Confluence specs, breaking API keywords).
  • Automated Remediation & Auditing: One-click trigger that generates and posts structured Atlassian Document Format (ADF) error panels directly into the Jira comment thread for release auditing.

How we built it

  • Framework: Atlassian Forge UI Kit (@forge/react) rendering natively inside Jira issue activity tabs.
  • Backend Architecture: Custom Node.js 22.x resolvers using @forge/api and @forge/bridge to securely interface with Atlassian Cloud endpoints.
  • APIs & Scopes: Jira REST API v3 for issue inspection and ADF comment writing (read:jira-work, write:jira-work), plus Confluence content summary integration (read:confluence-content.summary).

Challenges we ran into

  • ADF Comment Formatting: Designing robust, multi-node Atlassian Document Format JSON objects for automated error panel rendering.
  • Cloud Security Scopes: Configuring precise scope declarations in manifest.yml and managing production site authorization across Jira and Confluence environments.

Accomplishments that we're proud of

  • Delivering a fully functional, zero-latency risk evaluation dashboard running natively on Atlassian Cloud infrastructure without external servers.
  • Achieving complete end-to-end integration from API signal processing to automated Jira comment auditing.

What we learned

Deepened our knowledge of Atlassian Forge UI Kit components, custom resolver hooks, security scope management, and automated ADF payload construction.

What's next for ContextGuard AI

  • Deep integration with Bitbucket pull requests for real-time code change analysis.
  • Machine learning models trained on historical post-mortems to predict release incident probability.

Built With

Share this project:

Updates

posted an update —

Project Update: Automated Atlassian Document Format (ADF) Auditing

A key highlight of ContextGuard AI is automated release remediation. When high-risk signals are detected—such as unlinked specs or critical priority tags—the engine constructs structured Atlassian Document Format (ADF) JSON objects and posts formatted audit panels straight into the Jira issue thread for team sign-off.

See it in action in our demo video: https://youtu.be/qiEQyo1apy4

Submission history