Arbor

Arbor provides deterministic analysis of pull request (PR) blast radius, helping developers understand the actual impact of code changes before merging. By parsing the repository, building a call graph, and walking through changed paths, Arbor identifies which parts of the codebase a PR affects, aiming to catch issues that might otherwise go unnoticed. The tool is designed for software teams and developers seeking to assess the risk and reach of code modifications during the review process.

Arbor offers two primary modes of operation. Users can install its GitHub App for full graph analysis, enabling automatic graph walks on every PR and generating PR comments, dashboard runs, and repository-specific policies. Alternatively, for a quicker assessment without installation, Arbor allows users to paste a public PR diff or URL directly into the browser for a heuristic preview. This in-browser preview uses filename and keyword heuristics rather than the full repository call graph, providing a fast signal check with no signup or code storage required. Both methods generate shareable risk cards that summarize high-risk changes and suggest checks to perform before merging, such as adding regression tests for the riskiest branches or confirming backward compatibility of database migrations.

The platform emphasizes privacy, stating that no code is stored during heuristic analysis, and highlights its utility in collaborative review settings by making risk cards easily shareable in review threads, chat platforms, or agent prompts. Agent handoff features allow scoped repair prompts to be sent to external agents, focusing on patching risky branches and ensuring regression tests pass. Arbor also provides detailed guidance on review steps, such as running targeted tests, checking configuration changes, and verifying API response structures, to help teams catch subtle or high-impact issues before code is merged.

Arbor is available as a free tool for both public and private repositories. Its approach prioritizes earning user trust by allowing teams to try its analysis features with minimal setup before opting for deeper integration via the GitHub App.