--- title: CI/CD via agent tools: Skills, CLIs, MCP, and more with... description: A conversation with Semaphore's engineering and product leads walks through how the CI/CD platform rebuilt itself around AI coding agents over the past year.... canonical: https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore | daily.dev og:description: A conversation with Semaphore's engineering and product leads walks through how the CI/CD platform rebuilt itself around AI coding agents over the past year.... og:url: https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc og:image: https://api.daily.dev/og/posts/hwZc4BFjc.png og:image:alt: CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore 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. # CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore **[Bret Fisher](https://daily.dev/sources/bretfisher)** · 77 min read · 5 upvotes · 0 comments ## Summary A conversation with Semaphore's engineering and product leads walks through how the CI/CD platform rebuilt itself around AI coding agents over the past year. They open sourced their platform, built a CLI/MCP hybrid tool called SemAI with skills for tasks like flaky test fixing and Jenkins migration, and introduced ephemeral 'test boxes' that let agents run and debug CI jobs without needing a git push first. Topics covered include their runner benchmarks against competitors, self-hosted vs hosted runners, handling the surge of CVEs, contributor growth after going open source, and why agents now make CI migrations (e.g., off Jenkins) much more tractable than before. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are Semaphore test boxes and how do they let an AI agent fix failing CI pipelines? Test boxes are ephemeral CI machines that a developer or coding agent can spawn and SSH into directly, separate from the normal git-push-triggered pipeline. An agent can try multiple fixes on the box, run tests repeatedly, and iterate until the pipeline is green, without needing to commit and push code for every attempt, which is especially useful for large or environment-dependent tests. _Track how CI platforms like Semaphore support agent-driven test-and-fix loops via daily.dev._ ### How does Semaphore's SemAI tool combine a CLI and MCP server for AI coding agents? SemAI is built on Cobra and Viper so the same underlying commands power both a JSON-native CLI and a local MCP server, giving agents like Claude Code or Codex the choice of interface depending on the task. It ships as a plugin with descriptions, examples, and dozens of task-specific skills (such as migrating from Jenkins or fixing flaky tests) that auto-update when installed. _Compare CLI-plus-MCP tooling approaches for agent-driven CI/CD on daily.dev when evaluating platforms._ ### How can I restrict an AI coding agent's permissions when it interacts with a CI/CD platform like Semaphore? Semaphore supports creating multiple service accounts per project, each with its own token and role (reader, admin, etc.), so an agent can be scoped to only read pipeline status or only operate on a staging project while being blocked from production deployments or destructive actions. Secrets can also be attached per-project so an agent's ephemeral machine only has access to what that specific task needs. _Look into scoped service-account patterns for sandboxing agents in CI/CD via daily.dev._ ## Similar posts on daily.dev - [Why coding agents will break your CI/CD pipeline \(and how to fix it\)](https://daily.dev/posts/why-coding-agents-will-break-your-ci-cd-pipeline-and-how-to-fix-it--bikdwqfmf) · The New Stack · 1 upvotes · 0 comments - [Continuous AI in practice: What developers can automate today with agentic CI](https://daily.dev/posts/continuous-ai-in-practice-what-developers-can-automate-today-with-agentic-ci-bq4ihvy3k) · GitHub Blog · 2 upvotes · 0 comments - [SD Times News in Brief](https://daily.dev/posts/sd-times-news-in-brief-upqggjfrx) · SD Times · 0 upvotes · 0 comments - [The Dawn of Self-Improving Engineering: How AgentFactory Internalizes the CI/CD Lifecycle](https://daily.dev/posts/the-dawn-of-self-improving-engineering-how-agentfactory-internalizes-the-ci-cd-lifecycle-yjockeefl) · C\# Corner · 1 upvotes · 0 comments --- Tags: [#devops](https://daily.dev/tags/devops), [#ai-agents](https://daily.dev/tags/ai-agents), [#cicd](https://daily.dev/tags/cicd), [#mcp](https://daily.dev/tags/mcp) [View this post on daily.dev](https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc) ```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":"CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore","url":"https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc"},"datePublished":"2026-08-13T11:09:13.628Z","dateModified":"2026-09-14T06:30:08.386Z","description":"A conversation with Semaphore's engineering and product leads walks through how the CI/CD platform rebuilt itself around AI coding agents over the past year....","image":"https://i.ytimg.com/vi/HzXAdtVrW70/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/HzXAdtVrW70/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Bret Fisher","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":"Bret Fisher","logo":"https://media.daily.dev/image/upload/s--8RH_C-vW--/f_auto,q_auto/v1780213225/logos/bretfisher?_a=BAMAMiWQ0","url":"https://daily.dev/sources/bretfisher"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":5},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"devops,ai-agents,cicd,mcp","timeRequired":"PT77M","video":{"@type":"VideoObject","name":"CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore","description":"A conversation with Semaphore's engineering and product leads walks through how the CI/CD platform rebuilt itself around AI coding agents over the past year....","thumbnailUrl":"https://i.ytimg.com/vi/HzXAdtVrW70/sddefault.jpg","uploadDate":"2026-08-13T11:09:13.628Z","duration":"PT77M","url":"https://api.daily.dev/r/hwZc4BFjc","embedUrl":"https://www.youtube.com/embed/HzXAdtVrW70"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Bret Fisher","item":"https://daily.dev/sources/bretfisher"},{"@type":"ListItem","position":3,"name":"CI/CD via agent tools: Skills, CLIs, MCP, and more with Semaphore"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/ci-cd-via-agent-tools-skills-clis-mcp-and-more-with-semaphore-hwzc4bfjc#faq","mainEntity":[{"@type":"Question","name":"What are Semaphore test boxes and how do they let an AI agent fix failing CI pipelines?","acceptedAnswer":{"@type":"Answer","text":"Test boxes are ephemeral CI machines that a developer or coding agent can spawn and SSH into directly, separate from the normal git-push-triggered pipeline. An agent can try multiple fixes on the box, run tests repeatedly, and iterate until the pipeline is green, without needing to commit and push code for every attempt, which is especially useful for large or environment-dependent tests. Track how CI platforms like Semaphore support agent-driven test-and-fix loops via daily.dev."}},{"@type":"Question","name":"How does Semaphore's SemAI tool combine a CLI and MCP server for AI coding agents?","acceptedAnswer":{"@type":"Answer","text":"SemAI is built on Cobra and Viper so the same underlying commands power both a JSON-native CLI and a local MCP server, giving agents like Claude Code or Codex the choice of interface depending on the task. It ships as a plugin with descriptions, examples, and dozens of task-specific skills (such as migrating from Jenkins or fixing flaky tests) that auto-update when installed. Compare CLI-plus-MCP tooling approaches for agent-driven CI/CD on daily.dev when evaluating platforms."}},{"@type":"Question","name":"How can I restrict an AI coding agent's permissions when it interacts with a CI/CD platform like Semaphore?","acceptedAnswer":{"@type":"Answer","text":"Semaphore supports creating multiple service accounts per project, each with its own token and role (reader, admin, etc.), so an agent can be scoped to only read pipeline status or only operate on a staging project while being blocked from production deployments or destructive actions. Secrets can also be attached per-project so an agent's ephemeral machine only has access to what that specific task needs. Look into scoped service-account patterns for sandboxing agents in CI/CD via daily.dev."}}]} ```