SonarQube CLI 1.12.0 Adds Architecture Findings, Secret Redaction, and Simpler Project Administration
SonarQube CLI 1.12.0 adds architecture findings for SonarQube Cloud, redacts secrets in selected AI-agent workflows, and simplifies repository imports, project linking, and authentication management.
SonarQube CLI 1.12.0 brings architecture and security checks to the developer terminal
SonarQube CLI 1.12.0 is now available. The release adds an architecture findings command for SonarQube Cloud, changes how some detected secrets are handled, and lets teams handle more project administration without opening the web interface.
QA and software quality teams can now query architecture violations alongside issues, coverage, duplication, and dependency risks. That brings structural checks into local workflows and agent-assisted reviews, although architecture findings are currently limited to SonarQube Cloud.
Architecture findings become accessible from the CLI
The new sonar arch findings command reads architecture deviations, flaws, and smells from existing analysis results. Version 1.12.0 also adds filtering and pagination to architecture dependency views, so teams can narrow results by source and destination subtrees instead of reviewing every dependency relationship at once.
That is useful when a codebase has outgrown a single, easily digestible architecture report. A test or quality engineer investigating a change can query violations in the modules affected by a feature, while developers can focus on the boundaries tied to their work. The command does not replace full architecture governance, but it puts those findings into repeatable terminal-based checks.
Secret detection now supports redaction in selected agent workflows
The release changes how SonarQube hooks respond to some detected secrets. Previously, a detected secret could stop an agent session. In version 1.12.0, secrets found in prompts submitted to GitHub Copilot or files read by Codex can be masked, and the agent is told that redaction occurred. Content that cannot be safely masked remains blocked.
This gives AI-assisted development a more practical safety model. A hard stop is still appropriate when exposure cannot be controlled, while redaction can keep one accidental credential from ending an otherwise useful session. Teams should treat it as a safeguard, not a substitute for secret rotation, repository scanning, access controls, or shell-output review. The release notes limit the new behavior to particular Copilot and Codex hooks, so organizations should verify their own agent setup before depending on it.
Repository imports and project linking get simpler
Administrators running SonarQube Server can use sonar import to import repositories from GitHub, GitLab, Bitbucket Cloud, Bitbucket Server, and Azure DevOps. For teams standardizing analysis across several development groups, this moves initial project setup into the CLI instead of requiring repeated manual configuration.
The sonar link command now shows an interactive project picker when run without a project key. Rather than failing while the user searches for the key, it presents available projects for selection. Authentication cleanup is more targeted as well: sonar auth logout --server <url> removes the saved connection and its keychain entry for a specific server.
What QA teams should test after upgrading
- Run architecture queries against representative projects and confirm that filtering returns the expected source and destination subtrees.
- Test secret handling in the Copilot and Codex workflows used by your developers, including prompts, file reads, and content that should remain blocked.
- Validate repository imports on each supported source-control platform used by your organization.
- Check that project linking works for users with multiple SonarQube projects and that server-specific logout removes only the intended credentials.
- Review scripts that parse CLI output, since new commands and interactive behavior may affect automation.
SonarQube CLI 1.12.0 does not execute tests, but it expands the quality signals available before code reaches CI or pull-request review. For teams governing AI development workflows, architecture findings, agent-aware secret handling, and simpler project administration make the command line a more capable quality-control surface.
Looking for tools to match this read?
Jump into our independent side-by-side comparisons.
More in Tool Updates
.NET 11 Released: What QA Teams Need to Check in the New dotnet test
.NET 11 adds run-level test policies, clearer CI output, mobile test support, traversal projects, and new Microsoft.Testing.Platform options. QA teams should test these changes before upgrading.
GitHub Actions Secret Scanning Now Checks Workflow Logs: What QA Teams Need to Know
GitHub's public preview scans completed Actions workflow logs for leaked credentials. QA teams need to understand what it covers, how to enable it, and where it falls short before relying on it instead of disciplined test-pipeline hygiene.
Karate 2.1.4 Tightens Mock Defaults, Adds Request Limits, and Improves JavaScript Compatibility
Karate 2.1.4 was released October 22, with Maven Central artifacts published October 23. The update changes mock-server defaults, adds configurable request limits, and fixes several JavaScript compatibility issues.