detecting-existing-threads

Duplicate comments on pull requests occur when reviewers post inline findings without checking existing threads. Detecting existing threads before posting allows you to reuse matching discussions or update them instead of creating redundant comments, keeping PR feedback organized and reducing notification noise.

Detecting review threads involves fetching open and resolved inline comments from a PR, parsing them into a structured format, and comparing new findings against existing threads by location, content, and severity. This skill retrieves threads via gh CLI and a wrapper script, then applies exact, nearby, and content-based matching rules.

Yes. Automate detection by running this skill before posting inline findings. It fetches and parses existing PR threads, identifies matches based on line location and content similarity, and provides structured output showing which findings match existing threads and which require new comments.

You need the GitHub CLI (gh) installed and authenticated, access to the target pull request, and the get-review-threads.sh script to retrieve thread data. The skill resolves the PR number from environment variables, GitHub URLs, or local input, then queries existing comments and resolved threads.

Thread matching compares new findings against existing threads using exact location matching (same line), nearby matching (adjacent lines), and content-based matching (similar findings). Each match is mapped to severity levels and post-action rules, determining whether to reuse, update, or create new threads.

This skill detects both open (unresolved) and resolved inline review threads from PR comments. It fetches general PR-level comments and line-specific inline comments, unifying them into a single JSON structure for comparison against planned findings before posting new reviews.