More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    On-Call Runbook Builder: A Runbook from Your Repo Files

    1

    **What you get each run** - `runbook.md`: ten fixed sections, from "Service at a glance" to "Ask the team".

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    On-Call Runbook Builder: A Runbook from Your Repo Files

    On-Call Runbook Builder: A Runbook from Your Repo Files

    Example session with this skill installed

    We just took over this repo. Write the on-call runbook from its Compose file and Makefile, and list what we still need to ask the team.

    • Read your context and instructions
    • Compiled the on-call runbook builder

    $ python3 scripts/service_scan.py . --out runbook/facts.json
    summary: 51 facts from 6 files; 0 files not understood. Read from files only: nothing was run.

    Wrote runbook/runbook.md and runbook/runbook.html.

    $ python3 scripts/runbook_check.py runbook/runbook.md --repo .
    summary: 0 FIX, 0 ADVICE in runbook.md (12 steps, 10 commands, 9 open questions).

    Open questions for the team (from the runbook, not invented):

    • Who is on call for this service, and how are they reached out of hours?
    • What does /healthz check: only the process, or the database and Redis too?
    • Is there a database backup, where is it, and who may restore it?
    0 FIX, 0 ADVICE in runbook.md

    Connects securely to your tools. The creator never sees your data.

    About this skill

    English

    For developers, freelancers and agencies who run a service, or hand one over to a client or a new teammate.

    Point your agent at the service's repository. You get back an on-call runbook written from its own files: what runs, how to see its state, how to restart it, where the logs are, how a deploy and a rollback happen, and what to check first when it breaks.

    What you get each run

    • runbook.md: ten fixed sections, from "Service at a glance" to "Ask the team". Every command sits in a step that says what it affects, what you should see, and what to do if it fails.
    • runbook.html: the same runbook as one printable page, with read-only steps, restarts and production changes marked apart.
    • facts.json: what the scan found in your files, each fact with its file and line.

    Evidence for every command

    Each command names the file and line it came from, or says it is a standard command of a tool your files show in use. A bundled scan (service_scan.py, Python standard library) reads Docker Compose files, Dockerfiles, package.json scripts, Makefiles, Procfiles, env example files and GitHub Actions workflows. It only reads: it runs nothing, makes no network calls and never copies the value of an environment variable.

    Nothing invented

    Hosts, dashboards, alerts, contacts and backups are rarely in a repository. The runbook does not guess them: they become questions under "Ask the team", and the runbook is updated when you answer.

    How quality is kept

    runbook_check.py lints the runbook: all sections present, every command in a step with its effect, expected result and next move, every cited file and line real, every make target, npm script and Compose service found by the scan, and no command that removes volumes, drops data or rewrites history. The agent fixes and runs it again until no FIX is left. It checks structure and evidence, not whether a step is right for your production: walk through the runbook with the team once before relying on it.

    Arabic

    The runbook can be written in Arabic with a right-to-left page; commands, file names and variable names stay as they are.

    In the download: SKILL.md, the format reference, the three scripts, and a complete example so you can see the output before you buy: a small service repository (Compose, Dockerfile, Makefile, deploy workflow) and the English and Arabic runbooks made from it.

    بالعربي

    للمبرمجين والمستقلين والوكالات التي تشغّل خدمة، أو تسلّمها لعميل أو لزميل جديد.

    وجّه وكيلك إلى مستودع الخدمة. تستلم دليل مناوبة مكتوبًا من ملفاتها نفسها: ما الذي يعمل، وكيف تعرف حالته، وكيف تعيد تشغيله، وأين السجلات، وكيف يتم النشر والتراجع، وما أول ما تفحصه عند العطل.

    ماذا تستلم في كل مرة

    • ملف runbook.md: عشرة أقسام ثابتة، من «نظرة سريعة على الخدمة» إلى «اسأل الفريق». كل أمر داخل خطوة تذكر أثره، وما المتوقع أن تراه، وماذا تفعل إن لم ينجح.
    • صفحة runbook.html: الدليل نفسه كصفحة واحدة قابلة للطباعة، مع تمييز خطوات القراءة فقط عن إعادة التشغيل وعن تغيير الإنتاج.
    • ملف facts.json: ما وجده الفحص في ملفاتك، وكل معلومة مع ملفها وسطرها.

    دليل لكل أمر

    كل أمر يذكر الملف والسطر الذي جاء منه، أو يذكر أنه أمر قياسي لأداة تستعملها ملفاتك. أداة الفحص المرفقة service_scan.py (بايثون فقط) تقرأ ملفات Docker Compose وDockerfile وسكربتات package.json وMakefile وProcfile وأمثلة ملفات env وGitHub Actions. تقرأ فقط: لا تشغّل شيئًا، ولا تتصل بالشبكة، ولا تنسخ قيمة أي متغير.

    لا شيء مخترع

    أسماء الخوادم ولوحات المراقبة والتنبيهات وجهات الاتصال والنسخ الاحتياطية قلّما تكون في المستودع. الدليل لا يخمّنها: تصير أسئلة في قسم «اسأل الفريق»، ويُحدَّث الدليل عندما تجيب.

    كيف نحافظ على الجودة

    أداة runbook_check.py تدقق الدليل: الأقسام كلها موجودة، وكل أمر داخل خطوة فيها الأثر والمتوقع والخطوة التالية، وكل ملف وسطر مذكور موجود فعلًا، وكل هدف make وسكربت npm وخدمة Compose وجدها الفحص، ولا يوجد أمر يحذف وحدات تخزين أو بيانات أو يعيد كتابة التاريخ. يصلح الوكيل ويعيد الفحص حتى لا يبقى أي FIX. الأداة تفحص البنية والدليل، لا صحة الخطوة لبيئة إنتاجك: راجعوا الدليل مع الفريق مرة قبل الاعتماد عليه.

    العربية

    يمكن كتابة الدليل بالعربي مع صفحة من اليمين إلى اليسار، وتبقى الأوامر وأسماء الملفات والمتغيرات كما هي.

    داخل الملف: ملف SKILL.md، ومرجع الصيغة، والسكربتات الثلاثة، ومثال كامل لترى النتيجة قبل الشراء: مستودع خدمة صغير (Compose وDockerfile وMakefile ومسار نشر) والدليلان الإنجليزي والعربي المصنوعان منه.

    Requires Python 3. Skips Kubernetes, Helm, and Terraform files. Does not test commands against live environments. No values are extracted from env files, only keys.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 4 days ago

    Requires

    Googleapis
    Gstatic
    python3 (3.9 or newer)
    Python

    Permissions

    Terminal / Shell
    Network Access
    Read Files
    Write Files

    Allowed Hosts

    fonts.googleapis.com
    fonts.gstatic.com

    Reads the repository's configuration files, writes runbook.md, runbook.html and facts.json in your project, and runs its local Python scripts (service_scan.py, runbook_check.py, make_page.py). The scripts run none of your service's commands, make no network calls and never copy the value of an environment variable. يقرأ ملفات الإعداد ويكتب ملفات الدليل فقط.

    Listed4 days ago

    What's inside

    Frequently Asked Questions

    No. The scripts only read files in the repository. They run nothing and make no network calls. لا، السكربتات تقرأ ملفات المستودع فقط.

    Docker Compose files, Dockerfiles, package.json scripts, Makefiles, Procfiles, env example files (names only) and GitHub Actions workflows. Kubernetes manifests, Helm charts, systemd units and Terraform are listed as found, and the agent reads them by hand if they matter. ملفات Compose وDockerfile وpackage.json وMakefile وProcfile وأمثلة env وGitHub Actions.

    Treat it as a first draft. Every command cites its file and line so you can check it fast, and the open questions are listed. Walk through it with the team once before relying on it. اعتبره مسودة أولى؛ كل أمر معه مصدره، وراجعه مع الفريق مرة قبل الاعتماد عليه.

    Hosts, dashboards, alerts, contacts and backups are never invented. They become questions under "Ask the team", and the runbook is updated when you answer them. لا يُخترع شيء؛ تصير أسئلة للفريق.

    No. Commands that remove volumes, drop data or rewrite history are refused by the checker. Steps that restart something or change production say so in their first line. لا، وأداة الفحص ترفضها.

    Yes, as an added option: Arabic text with a right-to-left page, while commands and file names stay as they are. نعم، والأوامر وأسماء الملفات تبقى كما هي.

    New in DevOps & Cloud