--- title: Triaging open source with agents | daily.dev description: Bootstrap maintainers built an internal tool called twbs/triage that uses Claude Code agents to analyze every open issue and pull request, draft responses, and... canonical: https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Triaging open source with agents | daily.dev og:description: Bootstrap maintainers built an internal tool called twbs/triage that uses Claude Code agents to analyze every open issue and pull request, draft responses, and... og:url: https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx og:image: https://api.daily.dev/og/posts/h3PyRqYSx.png og:image:alt: Triaging open source with agents og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Triaging open source with agents **[Bootstrap](https://daily.dev/sources/bootstrap)** · 8 min read · 0 upvotes · 0 comments ## Summary Bootstrap maintainers built an internal tool called twbs/triage that uses Claude Code agents to analyze every open issue and pull request, draft responses, and prepare fix branches for human review. Each item moves through sync, analyze, prepare, store, and decide steps, gets an ease score from 0-10, and lands in a dashboard sorted by what needs attention. The first pass found nine issues already fixed, surfaced a buried but correct carousel fix, cleaned up GitHub labels across 198 items, and tracked individual unanswered points within long Discussions threads. Agents treat all contributor input as untrusted data, never run contributor code, and cannot post or merge on the main repo—maintainers still make every decision and write every response. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I use AI agents to triage a large backlog of open source GitHub issues and pull requests? Bootstrap built an internal tool called twbs/triage using Claude Code agents that follow five steps: sync open issues and PRs read-only, analyze the thread and reproduce the problem, prepare a drafted analysis, response, and fix branches, store everything in per-item folders, and let a maintainer decide. Each item gets an ease score from 0-10 indicating how much human work remains, from pasting a drafted comment to making a real scope decision. _Maintainers tracking agent-assisted workflows for their own repos can follow similar tooling patterns on daily.dev._ ### How does Bootstrap prevent AI agents from being manipulated by malicious contributor input during automated issue triage? Bootstrap treats everything a contributor submits as untrusted data rather than instructions. Agents never check out, install, or run contributor code, relying instead on diffs and GitHub CI as evidence; they never open links from an issue and instead build reproductions from scratch against Bootstrap's own dist files. Agents also cannot comment, push, or open pull requests on the main repo, except for applying labels from a fixed allowlist. _Teams designing agent-based automation around untrusted user input can compare safeguard approaches on daily.dev._ ### What results did Bootstrap get from running AI agents over its entire open issue backlog? A single pass by Claude Code agents over roughly 230 open issues and pull requests found nine items already fixed but never closed, surfaced a correct carousel touch-handling fix that had waited since July without review, and applied 410 labels across 198 open items after standardizing the label taxonomy. The process also broke down long GitHub Discussions threads, finding 17 still-open points within one 28-point feedback thread. _Anyone sizing up what AI-assisted triage can realistically deliver can follow real-world results like this on daily.dev._ ## Similar posts on daily.dev - [Copilot Triage: AI Issue and Pull Request Triage for GitHub](https://daily.dev/posts/copilot-triage-ai-issue-and-pull-request-triage-for-github-dwmywekcp) · RUBYLAND · 0 upvotes · 0 comments - [Meet Repro-Bot, our GitHub issue triage agent](https://daily.dev/posts/meet-repro-bot-our-github-issue-triage-agent-wifyv1fxp) · Metabase · 1 upvotes · 0 comments - [Github Integrates AI to Improve Accessibility Issue Management and Automate Feedback Triage](https://daily.dev/posts/github-integrates-ai-to-improve-accessibility-issue-management-and-automate-feedback-triage-3vu0hjwbk) · InfoQ · 0 upvotes · 0 comments --- Tags: [#open-source](https://daily.dev/tags/open-source), [#ai-agents](https://daily.dev/tags/ai-agents), [#github](https://daily.dev/tags/github), [#claude-code](https://daily.dev/tags/claude-code), [#bootstrap-css](https://daily.dev/tags/bootstrap-css) [View this post on daily.dev](https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Triaging open source with agents","url":"https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx"},"datePublished":"2026-10-15T15:38:09.917Z","dateModified":"2026-10-15T15:45:31.752Z","description":"Bootstrap maintainers built an internal tool called twbs/triage that uses Claude Code agents to analyze every open issue and pull request, draft responses, and...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/26d08898ac2f5b859109968a4822eeb9?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/26d08898ac2f5b859109968a4822eeb9?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Bootstrap","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Bootstrap","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/ea4f1ba8a2d6493e853e77a6002c9c00","url":"https://daily.dev/sources/bootstrap"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"open-source,ai-agents,github,claude-code,bootstrap-css","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Bootstrap","item":"https://daily.dev/sources/bootstrap"},{"@type":"ListItem","position":3,"name":"Triaging open source with agents"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/triaging-open-source-with-agents-h3pyrqysx#faq","mainEntity":[{"@type":"Question","name":"How can I use AI agents to triage a large backlog of open source GitHub issues and pull requests?","acceptedAnswer":{"@type":"Answer","text":"Bootstrap built an internal tool called twbs/triage using Claude Code agents that follow five steps: sync open issues and PRs read-only, analyze the thread and reproduce the problem, prepare a drafted analysis, response, and fix branches, store everything in per-item folders, and let a maintainer decide. Each item gets an ease score from 0-10 indicating how much human work remains, from pasting a drafted comment to making a real scope decision. Maintainers tracking agent-assisted workflows for their own repos can follow similar tooling patterns on daily.dev."}},{"@type":"Question","name":"How does Bootstrap prevent AI agents from being manipulated by malicious contributor input during automated issue triage?","acceptedAnswer":{"@type":"Answer","text":"Bootstrap treats everything a contributor submits as untrusted data rather than instructions. Agents never check out, install, or run contributor code, relying instead on diffs and GitHub CI as evidence; they never open links from an issue and instead build reproductions from scratch against Bootstrap's own dist files. Agents also cannot comment, push, or open pull requests on the main repo, except for applying labels from a fixed allowlist. Teams designing agent-based automation around untrusted user input can compare safeguard approaches on daily.dev."}},{"@type":"Question","name":"What results did Bootstrap get from running AI agents over its entire open issue backlog?","acceptedAnswer":{"@type":"Answer","text":"A single pass by Claude Code agents over roughly 230 open issues and pull requests found nine items already fixed but never closed, surfaced a correct carousel touch-handling fix that had waited since July without review, and applied 410 labels across 198 open items after standardizing the label taxonomy. The process also broke down long GitHub Discussions threads, finding 17 still-open points within one 28-point feedback thread. Anyone sizing up what AI-assisted triage can realistically deliver can follow real-world results like this on daily.dev."}}]} ```