<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
  <title>StarSling Blog</title>
  <subtitle>Notes on faster CI, AI-native pipelines, and StarSling engineering.</subtitle>
  <link rel="self" type="application/atom+xml" href="https://starsling.dev/rss.xml"/>
  <link rel="alternate" type="text/html" href="https://starsling.dev/blog"/>
  <id>tag:starsling.dev,2026:blog</id>
  <updated>2026-11-17T00:00:00.000Z</updated>
  <icon>https://starsling.dev/images/logos/starsling-logo-light.svg</icon>
  <author>
    <name>StarSling</name>
    <email>hello@starsling.dev</email>
  </author>
  <entry>
    <title type="text">Announcing the CI Challenge</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/ci-challenge"/>
    <id>tag:starsling.dev,2026:blog/ci-challenge</id>
    <published>2026-11-17T00:00:00.000Z</published>
    <updated>2026-11-17T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">If StarSling can&apos;t save you at least 10% on CI in 30 days, we&apos;ll wire your organization $1,000.</summary>
    <content type="html">&lt;p&gt;StarSling is self-improving CI that&apos;s designed for agent-volume repos. You get fast GitHub Actions runners with AI agents that continuously improve your CI by sending you PRs that optimize your workflows and tests. Our &lt;a href=&quot;/blog/agent-loop&quot;&gt;CI harness&lt;/a&gt; has helped companies like &lt;a href=&quot;/customers/mastra&quot;&gt;Mastra get a 6x CI speedup&lt;/a&gt;.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/ci-challenge&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/ci-challenge-hero-d3234a98.jpg"/>
  </entry>
  <entry>
    <title type="text">Introducing Disks: files ready 10.5x faster</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/disks"/>
    <id>tag:starsling.dev,2026:blog/disks</id>
    <published>2026-11-13T00:00:00.000Z</published>
    <updated>2026-11-13T00:00:00.000Z</updated>
    <author><name>Daniel Worku</name></author>
    <summary type="text">StarSling Disks mount persistent filesystems into GitHub Actions jobs. Mastra&apos;s dependencies are ready in 11.5 seconds, down from 56, across 350,000+ jobs.</summary>
    <content type="html">&lt;p&gt;Introducing &lt;a href=&quot;https://docs.starsling.dev/runners/disks&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;&lt;strong&gt;Disks&lt;/strong&gt;&lt;/a&gt;: persistent filesystems mounted into &lt;a href=&quot;/products/runners&quot;&gt;StarSling Runners&lt;/a&gt; as ordinary directories. A job reads and writes its dependency store, build output, or dataset in place, and the next job finds it already there.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/disks&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/disks-hero-5bb03811-95f73598.jpg"/>
  </entry>
  <entry>
    <title type="text">Introducing CI Playbooks: 250+ GitHub Actions Lessons from Software Factories</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/introducing-ci-playbooks"/>
    <id>tag:starsling.dev,2026:blog/introducing-ci-playbooks</id>
    <published>2026-10-29T00:00:00.000Z</published>
    <updated>2026-10-29T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">CI Playbooks collect the CI techniques of ten companies running software factories: 253 techniques, each sourced, with an agent prompt to apply it.</summary>
    <content type="html">&lt;p&gt;Faster runners are one part of how the best engineering teams keep CI fast. The rest is changes to the workflows and tests themselves. We gathered these CI changes from ten companies running software factories, using their own posts, talks and pull requests: Anthropic, DoorDash, Linear, Mastra, OpenAI, Ramp, Shopify, Spotify, Stripe and Uber.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/introducing-ci-playbooks&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/ci-playbooks-index-og-8c14c7c8.jpg"/>
  </entry>
  <entry>
    <title type="text">OpenAI&apos;s CI playbook (for GitHub Actions)</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/openai-ci"/>
    <id>tag:starsling.dev,2026:blog/openai-ci</id>
    <published>2026-10-22T00:00:00.000Z</published>
    <updated>2026-10-22T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">OpenAI&apos;s Core Experimentation team cut median CI time from about 10 minutes to 4.5. Ten of its techniques, each with a copy-paste prompt for your agent.</summary>
    <content type="html">&lt;p&gt;Last week, OpenAI shared &lt;a href=&quot;https://developers.openai.com/blog/maintaining-software-with-the-gpt-6-family&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;how its Core Experimentation team uses Codex&lt;/a&gt; to maintain its software, from removing unused code and stale feature flags to speeding up CI.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/openai-ci&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/openai-ci-hero-a15621d3.jpg"/>
  </entry>
  <entry>
    <title type="text">Anthropic&apos;s CI playbook, as GitHub Actions prompts for your agent</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/anthropic-ci"/>
    <id>tag:starsling.dev,2026:blog/anthropic-ci</id>
    <published>2026-10-19T00:00:00.000Z</published>
    <updated>2026-10-19T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Anthropic&apos;s CI jobs grew 25x in six months. How to run only the tests a change affects in GitHub Actions, with a copy-paste prompt for your agent per technique.</summary>
    <content type="html">&lt;p&gt;Last month &lt;a href=&quot;https://claude.com/resources/articles/agentic-coding-is-straining-ci-heres-how-we-scaled-test-impact-analysis-at-anthropic&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;Anthropic explained how they scaled CI using test impact analysis&lt;/a&gt;: running only the tests a change can affect. Their engineers now ship 8x as much code per quarter as they did from 2021-2025, and their CI jobs grew 25x in six months. Their merges went from about 2,500 a day in May to more than 30,000 a day in September, &lt;a href=&quot;https://www.anthropic.com/engineering/redesigning-source-control&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;according to their source control team&lt;/a&gt;. The post explains why CI strains once agents write most of the code, and how they rebuilt their test selection service to keep up.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/anthropic-ci&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/anthropic-ci-hero-86bd0fbb.jpg"/>
  </entry>
  <entry>
    <title type="text">Introducing Adaptive Compute: up to 2.3x faster GitHub Actions</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/adaptive-compute"/>
    <id>tag:starsling.dev,2026:blog/adaptive-compute</id>
    <published>2026-10-14T00:00:00.000Z</published>
    <updated>2026-10-14T00:00:00.000Z</updated>
    <author><name>Daniel Worku</name></author>
    <summary type="text">StarSling Runners with Adaptive Compute ran up to 2.32x faster than GitHub-hosted runners, boosting CPU up to 25% and RAM up to 50% at no extra cost.</summary>
    <content type="html">&lt;p&gt;Your CI workload changes between jobs and even within individual steps. Its compute should too.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/adaptive-compute&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/afterburner-og-b15e523f.jpg"/>
  </entry>
  <entry>
    <title type="text">Linear&apos;s CI playbook, ready to hand to your agent</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/linear-ci"/>
    <id>tag:starsling.dev,2026:blog/linear-ci</id>
    <published>2026-10-13T00:00:00.000Z</published>
    <updated>2026-10-13T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Linear&apos;s 15 CI changes from their post, each with a copy-paste prompt that hands the move to your coding agent and names the ci-speedup pattern behind it.</summary>
    <content type="html">&lt;p&gt;A few weeks ago &lt;a href=&quot;https://linear.app/now/ci-bottleneck-reworked&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;Linear published a blog post&lt;/a&gt; about how they reworked their CI. Their optimizations led to a 15% faster test suite and 50% reduction in runner-time spent per test, while their test suite grew 4x in size. The post was popular &lt;a href=&quot;https://x.com/linear/status/2102098669889552711&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;on X&lt;/a&gt; and Hacker News, and has a lot of technical depth (with pretty graphics).&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/linear-ci&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/linear-ci-hero-2178b028.jpg"/>
  </entry>
  <entry>
    <title type="text">Ubuntu 26 generally available</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/ubuntu-26"/>
    <id>tag:starsling.dev,2026:blog/ubuntu-26</id>
    <published>2026-10-12T00:00:00.000Z</published>
    <updated>2026-10-12T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">GitHub moves ubuntu-latest to Ubuntu 26.04 starting Oct 19, 2026, and the switch can break builds. Use our agent prompt to check if your jobs are affected.</summary>
    <content type="html">&lt;p&gt;On September 17th, 2026 &lt;a href=&quot;https://github.blog/changelog/2026-09-17-ubuntu-26-generally-available-and-latest-migration/#ubuntu-latest-is-migrating-to-ubuntu-26-04&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;GitHub announced&lt;/a&gt; the &lt;code&gt;ubuntu-latest&lt;/code&gt; Actions runner label will migrate from Ubuntu 24.04 to Ubuntu 26.04 starting on October 19, 2026 which is one week away.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/ubuntu-26&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/ubuntu-26-runner-swap-fc1fe131.jpg"/>
  </entry>
  <entry>
    <title type="text">StarSling tops the GitHub Actions runner benchmarks</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/gha-benchmarks"/>
    <id>tag:starsling.dev,2026:blog/gha-benchmarks</id>
    <published>2026-10-05T00:00:00.000Z</published>
    <updated>2026-10-05T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">StarSling Runners finished first on the Rust and TypeScript builds in RunsOn&apos;s public GitHub Actions runner benchmark, across 9 providers and 89 runners.</summary>
    <content type="html">&lt;p&gt;&lt;a href=&quot;/products/runners&quot;&gt;StarSling Runners&lt;/a&gt; came in first on both the Rust and TypeScript builds in RunsOn&apos;s open &lt;a href=&quot;https://runs-on.com/benchmarks/runners/starsling/&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;GitHub Actions runner benchmark&lt;/a&gt;, which runs the same cold builds on 89 runners from 9 providers. The harness and every raw result are &lt;a href=&quot;https://github.com/runs-on-demo/benchmark-public/tree/33ed4431b26868d810939bea9315081f31bd0c74&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;public on GitHub&lt;/a&gt;.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/gha-benchmarks&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/tops-benchmarks-card-4d964fd7.jpg"/>
  </entry>
  <entry>
    <title type="text">Levels of CI</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/levels-of-ci"/>
    <id>tag:starsling.dev,2026:blog/levels-of-ci</id>
    <published>2026-09-30T00:00:00.000Z</published>
    <updated>2026-09-30T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Five levels of CI optimization, from the Actions cache to swapping runners, tuning workflows, switching CI engines, and replacing CI with agents.</summary>
    <content type="html">&lt;p&gt;Everyone is talking about CI these days. Anthropic&apos;s CI job volume &lt;a href=&quot;https://claude.com/blog/agentic-coding-is-straining-ci-heres-how-we-scaled-test-impact-analysis-at-anthropic&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;increased 25x over 6 months&lt;/a&gt; while their tests &lt;a href=&quot;https://x.com/addyosmani/status/2099577600159158765&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;grew 10x&lt;/a&gt;. Linear took up a major effort to &lt;a href=&quot;https://linear.app/now/ci-bottleneck-reworked&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;rework their CI&lt;/a&gt; after their test suite &lt;a href=&quot;https://x.com/moofeez/status/2102110907295993918&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;grew 3.7x&lt;/a&gt;. Meanwhile, more and more people are complaining about their GitHub Actions costs being out of control.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/levels-of-ci&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/levels-of-ci-hero-v2-a22aeb1d.jpg"/>
  </entry>
  <entry>
    <title type="text">Announcing Our $3M Pre-Seed Round and StarSling Review Runners</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/pre-seed"/>
    <id>tag:starsling.dev,2026:blog/pre-seed</id>
    <published>2026-09-22T00:00:00.000Z</published>
    <updated>2026-09-22T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">A $3M pre-seed from Bessemer Venture Partners, Y Combinator and five more funds to make shipping code agent-native, and Review Runners in private beta.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re announcing that StarSling has raised a &lt;strong&gt;$3 million pre-seed&lt;/strong&gt; round from Bessemer Venture Partners, Y Combinator, and several other amazing investors, and that we&apos;re launching &lt;a href=&quot;/products/review-runners&quot;&gt;&lt;strong&gt;StarSling Review Runners&lt;/strong&gt;&lt;/a&gt;.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/pre-seed&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/starsling-3m-preseed-review-runners-og-bfb66272.jpg"/>
  </entry>
  <entry>
    <title type="text">Announcing StarSling GPU Runners</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/announcing-starsling-gpu-runners"/>
    <id>tag:starsling.dev,2026:blog/announcing-starsling-gpu-runners</id>
    <published>2026-09-11T00:00:00.000Z</published>
    <updated>2026-09-11T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">GitHub Actions runners backed by NVIDIA GPUs: RTX PRO 6000, RTX 5090 and RTX 4090, with H100 coming. One line in runs-on. Contact us to get started.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re announcing &lt;a href=&quot;/products/gpu-runners&quot;&gt;&lt;strong&gt;StarSling GPU Runners&lt;/strong&gt;&lt;/a&gt;: GitHub Actions runners backed by NVIDIA GPUs. &lt;a href=&quot;https://tally.so/r/Y56E8v&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot; data-ph-capture-attribute-cta=&quot;primary&quot; data-ph-capture-attribute-cta-variant=&quot;gpu_access&quot; data-ph-capture-attribute-cta-location=&quot;blog_body&quot;&gt;Contact us to get started&lt;/a&gt;.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/announcing-starsling-gpu-runners&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/announcing-starsling-gpu-runners-og-3f4609d3.jpg"/>
  </entry>
  <entry>
    <title type="text">The agent loop that shipped a 5.4x speedup</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/agent-loop"/>
    <id>tag:starsling.dev,2026:blog/agent-loop</id>
    <published>2026-09-09T00:00:00.000Z</published>
    <updated>2026-09-09T00:00:00.000Z</updated>
    <author><name>Yunjia Zheng</name></author>
    <summary type="text">The job was green. The logs were clean. CI Harness v2 found the 11-month-old bottleneck anyway, and made that job 5.4x faster.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re announcing StarSling CI Harness v2. It is more intelligent, more
creative, and produces fewer false positives.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/agent-loop&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/agent-loop-og-cff536cf.jpg"/>
  </entry>
  <entry>
    <title type="text">Introducing sling: an agent-first CLI for your CI</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/introducing-sling-an-agent-first-cli-for-your-ci"/>
    <id>tag:starsling.dev,2026:blog/introducing-sling-an-agent-first-cli-for-your-ci</id>
    <published>2026-08-27T00:00:00.000Z</published>
    <updated>2026-08-27T00:00:00.000Z</updated>
    <author><name>Shreyansh Saurabh</name></author>
    <summary type="text">sling is the agent-first CLI for StarSling Runners. Diagnose a failed job, rank your slowest workflows, see CI costs, and read only the log lines that matter.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re releasing &lt;a href=&quot;/cli&quot;&gt;&lt;strong&gt;sling&lt;/strong&gt;&lt;/a&gt;, the official StarSling CLI. It&apos;s an
agent-first CLI for your CI.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/introducing-sling-an-agent-first-cli-for-your-ci&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/sling-blog-thumbnail-d7a12c07.gif"/>
  </entry>
  <entry>
    <title type="text">More GitHub Actions optimizations: 16 recipes your agent can apply</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/more-github-actions-optimizations-16-recipes-your-agent-can-apply"/>
    <id>tag:starsling.dev,2026:blog/more-github-actions-optimizations-16-recipes-your-agent-can-apply</id>
    <published>2026-08-24T00:00:00.000Z</published>
    <updated>2026-08-24T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Sixteen GitHub Actions optimization recipes with runnable YAML, tradeoffs, and verify steps, plus one JSON file your coding agent fetches to find the right one.</summary>
    <content type="html">&lt;p&gt;Last week we published &lt;a href=&quot;/blog/github-actions-optimizations-18-best-practices-guides-you-can-give-to-your-agent-to-read&quot;&gt;18 best practices guides&lt;/a&gt;: one page per practice, on what good CI looks like. They are written to be handed to a coding agent, and every one is served as plain markdown at the same URL plus &lt;code&gt;.md&lt;/code&gt;, so an agent reads the page without executing a line of JavaScript.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/more-github-actions-optimizations-16-recipes-your-agent-can-apply&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/github-actions-16-recipes-og-0b71ee1b.jpg"/>
  </entry>
  <entry>
    <title type="text">GitHub Actions optimizations: 18 best practices guides you can give to your agent to read</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/github-actions-optimizations-18-best-practices-guides-you-can-give-to-your-agent-to-read"/>
    <id>tag:starsling.dev,2026:blog/github-actions-optimizations-18-best-practices-guides-you-can-give-to-your-agent-to-read</id>
    <published>2026-08-19T21:02:43.000Z</published>
    <updated>2026-08-19T21:02:43.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Eighteen free GitHub Actions best practices guides with copyable YAML, verify steps, and markdown mirrors your coding agent can read and apply to your repo.</summary>
    <content type="html">&lt;p&gt;Over the last few months we shipped three &lt;a href=&quot;https://github.com/starslingdev/skills&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;open-source&lt;/a&gt; agent skills that read your GitHub Actions workflows and fix what is wrong with them: &lt;a href=&quot;/skills/ci-speedup&quot;&gt;ci-speedup&lt;/a&gt; for what is slowing CI down, &lt;a href=&quot;/skills/ci-secure&quot;&gt;ci-secure&lt;/a&gt; for the critical attack vectors, and &lt;a href=&quot;/skills/ci-score&quot;&gt;ci-score&lt;/a&gt; for how your configuration grades overall. Every finding those skills report rests on a practice: a specific thing a workflow should do, and a specific reason it should.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/github-actions-optimizations-18-best-practices-guides-you-can-give-to-your-agent-to-read&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/jpeg" href="https://starsling.dev/static/github-actions-best-practices-18-guides-og-v2-1a6f6450.jpg"/>
  </entry>
  <entry>
    <title type="text">Introducing ci-secure: an AI coding agent skill that closes the critical attack vectors in your GitHub Actions</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/introducing-ci-secure-an-ai-coding-agent-skill-that-closes-the-critical-attack-vectors-in-your-github-actions"/>
    <id>tag:starsling.dev,2026:blog/introducing-ci-secure-an-ai-coding-agent-skill-that-closes-the-critical-attack-vectors-in-your-github-actions</id>
    <published>2026-08-13T17:01:34.000Z</published>
    <updated>2026-08-13T17:01:34.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Scan your GitHub Actions for the ten critical CI/CD attack vectors. ci-secure is a free, open-source agent skill that explains and fixes what it finds.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re releasing &lt;a href=&quot;/skills/ci-secure&quot;&gt;ci-secure&lt;/a&gt;, a free and &lt;a href=&quot;https://github.com/starslingdev/skills&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;open-source&lt;/a&gt; agent skill that scans your GitHub Actions workflows for the ten critical CI/CD attack vectors and hands your coding agent a fix for every one it finds. It works with all coding agents supported by &lt;a href=&quot;https://github.com/vercel-labs/skills#supported-agents&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;Vercel&apos;s Skills CLI&lt;/a&gt;, including OpenCode, Claude Code, Codex, and Cursor.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/introducing-ci-secure-an-ai-coding-agent-skill-that-closes-the-critical-attack-vectors-in-your-github-actions&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/ci-secure-blog-thumbnail-f0c66b6f.gif"/>
  </entry>
  <entry>
    <title type="text">Results Explorer and Updates for High Performance Sandbox Benchmarks</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/results-explorer-and-updates-for-high-performance-sandbox-benchmarks"/>
    <id>tag:starsling.dev,2026:blog/results-explorer-and-updates-for-high-performance-sandbox-benchmarks</id>
    <published>2026-08-07T00:00:00.000Z</published>
    <updated>2026-08-07T00:00:00.000Z</updated>
    <author><name>Daniel Worku</name></author>
    <summary type="text">Our sandbox benchmarks now have an interactive results explorer and a refreshed run covering 11 environments. Pick any metric and percentile, share the view.</summary>
    <content type="html">&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; We&apos;ve added an interactive results explorer to &lt;a href=&quot;/hpc-sandbox-benchmarks&quot;&gt;High Performance Sandbox Benchmarks&lt;/a&gt; and refreshed the underlying run. It now covers 11 sandbox environments, up from 6. The harness and every raw run are still open source at &lt;a href=&quot;https://github.com/starslingdev/hpc-sandbox-benchmarks&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;starslingdev/hpc-sandbox-benchmarks&lt;/a&gt;.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/results-explorer-and-updates-for-high-performance-sandbox-benchmarks&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/results-explorer-19ca7643.gif"/>
  </entry>
  <entry>
    <title type="text">Introducing ci-score: an AI coding agent skill that grades your GitHub Actions setup</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/introducing-ci-score-an-ai-coding-agent-skill-that-grades-your-github-actions-setup"/>
    <id>tag:starsling.dev,2026:blog/introducing-ci-score-an-ai-coding-agent-skill-that-grades-your-github-actions-setup</id>
    <published>2026-07-30T00:00:00.000Z</published>
    <updated>2026-07-30T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Check your GitHub Actions against best practices. ci-score is a free, open-source agent skill that grades your config and ranks a fix for every gap.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re releasing &lt;a href=&quot;/skills/ci-score&quot;&gt;ci-score&lt;/a&gt;, a free and &lt;a href=&quot;https://github.com/starslingdev/skills&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;open-source&lt;/a&gt; agent skill that grades your GitHub Actions configuration against a published rubric of best practices and hands your coding agent a ranked fix for every gap it finds. It works with all &lt;a href=&quot;https://github.com/vercel-labs/skills#supported-agents&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;73 coding agents currently supported by Vercel&apos;s Skills CLI&lt;/a&gt;, including OpenCode, Claude Code, Codex, and Cursor.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/introducing-ci-score-an-ai-coding-agent-skill-that-grades-your-github-actions-setup&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/ci-score-blog-thumbnail-710d5ec0.gif"/>
  </entry>
  <entry>
    <title type="text">Beyond Cold-Starts: Benchmarking Sandboxes for Real Workloads</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/beyond-cold-starts-benchmarking-sandboxes-for-real-workloads"/>
    <id>tag:starsling.dev,2026:blog/beyond-cold-starts-benchmarking-sandboxes-for-real-workloads</id>
    <published>2026-07-23T00:00:00.000Z</published>
    <updated>2026-07-23T00:00:00.000Z</updated>
    <author><name>Daniel Worku</name></author>
    <summary type="text">We benchmarked cloud sandbox providers (Blaxel, Daytona, E2B, Modal, Novita) on real CI workloads and open-sourced the results. Startup speed isn&apos;t the story.</summary>
    <content type="html">&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; After running 500K+ sandbox jobs in prod, we&apos;ve &lt;a href=&quot;https://github.com/starslingdev/hpc-sandbox-benchmarks&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;open-sourced benchmarks&lt;/a&gt; for cloud sandbox providers using real-world CI workloads. For our workloads, sandboxes from Blaxel and Daytona&apos;s microVMs have excellent performance.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/beyond-cold-starts-benchmarking-sandboxes-for-real-workloads&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/png" href="https://starsling.dev/static/sandbox-benchmarks-og-card-35df60bc.png"/>
  </entry>
  <entry>
    <title type="text">Introducing ci-speedup: an AI coding agent skill that analyzes and fixes your slow GitHub Actions</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/introducing-ci-speedup-an-ai-coding-agent-skill-that-analyzes-and-fixes-your-slow-github-actions"/>
    <id>tag:starsling.dev,2026:blog/introducing-ci-speedup-an-ai-coding-agent-skill-that-analyzes-and-fixes-your-slow-github-actions</id>
    <published>2026-07-21T00:00:00.000Z</published>
    <updated>2026-07-21T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">ci-speedup is a free, open-source agent skill that analyzes GitHub Actions runs, finds the blocking check, and gives your coding agent the evidence to fix it.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re releasing &lt;a href=&quot;/skills/ci-speedup&quot;&gt;ci-speedup&lt;/a&gt;, a free and &lt;a href=&quot;https://github.com/starslingdev/skills&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;open-source&lt;/a&gt; agent skill that finds the one thing actually slowing your pull requests down and gives your coding agent the evidence-backed prompt to fix it. It works with all &lt;a href=&quot;https://github.com/vercel-labs/skills#supported-agents&quot; target=&quot;_blank&quot; rel=&quot;noopener noreferrer&quot;&gt;73 coding agents currently supported by Vercel&apos;s Skills CLI&lt;/a&gt;, including OpenCode, Claude Code, Codex, and Cursor.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/introducing-ci-speedup-an-ai-coding-agent-skill-that-analyzes-and-fixes-your-slow-github-actions&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/ci-speedup-blog-thumbnail-ffe2d0cb.gif"/>
  </entry>
  <entry>
    <title type="text">StarSling Runners are now generally available</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/starsling-runners-are-now-generally-available"/>
    <id>tag:starsling.dev,2026:blog/starsling-runners-are-now-generally-available</id>
    <published>2026-06-29T00:00:00.000Z</published>
    <updated>2026-06-29T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">StarSling Runners are generally available: self-driving CI for GitHub Actions. Install the GitHub App and swap to starsling-ubuntu-26.04. 2,000 free minutes.</summary>
    <content type="html">&lt;p&gt;Today we&apos;re launching &lt;a href=&quot;/products/runners&quot;&gt;StarSling Runners: Self-Driving CI&lt;/a&gt; in general availability.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/starsling-runners-are-now-generally-available&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/gif" href="https://starsling.dev/static/starsling-runners-ga-launch-runner-swap-optimized-9dc684e7.gif"/>
  </entry>
  <entry>
    <title type="text">Announcing StarSling Runners: self-driving CI that optimizes itself</title>
    <link rel="alternate" type="text/html" href="https://starsling.dev/blog/announcing-starsling-runners-self-driving-ci-that-optimizes-itself"/>
    <id>tag:starsling.dev,2026:blog/announcing-starsling-runners-self-driving-ci-that-optimizes-itself</id>
    <published>2026-04-30T00:00:00.000Z</published>
    <updated>2026-04-30T00:00:00.000Z</updated>
    <author><name>Yonas Beshawred</name></author>
    <summary type="text">Fast GitHub Actions runners with AI agents that continuously ship optimization PRs. Customers like Better Auth and Mastra are seeing 82% faster CI.</summary>
    <content type="html">&lt;p&gt;&lt;strong&gt;Update - June 29, 2026:&lt;/strong&gt; StarSling Runners are now generally available. See the &lt;a href=&quot;https://starsling.dev/blog/starsling-runners-are-now-generally-available&quot;&gt;GA announcement&lt;/a&gt; for current availability, pricing, and AI optimization access. For new accounts, AI-powered optimization PRs are only available on paid plans and are not enabled by default.&lt;/p&gt;&lt;p&gt;&lt;a href=&quot;https://starsling.dev/blog/announcing-starsling-runners-self-driving-ci-that-optimizes-itself&quot;&gt;Read the full post on starsling.dev →&lt;/a&gt;&lt;/p&gt;</content>
    <link rel="enclosure" type="image/png" href="https://starsling.dev/static/announcing-starsling-runners-2784a465.png"/>
  </entry>
</feed>
