You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Overall impression: As a brand-new user, the docs are clear and well organized — the home page immediately explains what the tool does and points to "Get Started with CLI," and the Quick Start guide walks through a full working example (install → add sample workflow → wait for result → customize) with realistic time estimates and a visual example of the output.
Critical Issues Found
None found that block getting started. All three pages loaded correctly, no 404s, and the documented CLI commands (gh extension install github/gh-aw, gh aw add-wizard, gh aw compile, gh aw run) are shown with clear syntax.
One near-miss worth flagging: the sample-output image on the Quick Start page ("The report will look something like this:") initially reported naturalWidth/Height = 0 via script inspection before it scrolled into view — this turned out to be lazy-loading behavior, not an actual broken image (confirmed 200 response and full load after scroll), but a first-time reader skimming quickly could briefly wonder if the image is broken.
Confusing Areas
Jargon introduced without inline definition: the intro paragraph mentions "recurring agentic workflow" and "the stable-engine path" without explicitly defining "engine" (AI engine/runtime) before this point — a newcomer may not yet know that "engine" = Copilot/Claude/Codex/Gemini/Pi choice.
Five parallel setup guides collapsed into tabs: the "Configuring authentication" section has 5 tabs (Copilot/Claude/Codex/Gemini/Pi) but only the Copilot tab content was inspected by default — a first-time user has to know to click through tabs to find instructions for their chosen engine; it's not obvious from a glance which tab is "selected" by default relative to what was chosen in Step 2.
gh aw add-wizard walks through 6 sub-steps (check prerequisites, select engine, configure auth, add workflow, add skills, trigger run) listed in prose — a numbered list or short animated GIF of the actual interactive prompts would make expectations clearer before running the command for real.
What Worked Well
Home page nav is excellent: "Quick Start," "Create," "Gallery," "Docs," "FAQ" are all one click away, and the two CTA buttons ("Get Started with CLI" and "Creating Workflows") make the very next action obvious.
Quick Start prerequisites are listed as a clean checklist (AI account, repo, Actions enabled, gh CLI v2.0.0+, login, OS) — no assumed context.
Time estimates ("~10 minutes," "2–3 minutes for the run") set good expectations up front.
The CLI Commands page leads with a "Day-one commands" table that includes a "When to use" column — this is much more useful for a beginner than an alphabetical command reference, and a tip box explicitly says "New to gh aw? Start with the day-one commands... can be skipped for most users," which is reassuring.
Fallback install path (curl ... install-gh-aw.sh | bash) is provided for users hitting gh extension install auth issues — good anticipation of a common failure mode.
Version pinning instructions (gh extension install github/gh-aw@v0.1.0) are clearly explained for teams wanting reproducibility.
Recommendations
Quick wins:
Add a one-line inline definition or tooltip for "engine" the first time it's used on the Quick Start page (e.g., "engine — the AI provider/runtime that executes the workflow, e.g., GitHub Copilot, Claude, Codex, Gemini, or Pi").
Indicate visually which authentication tab (Copilot/Claude/Codex/Gemini/Pi) corresponds to the engine chosen in the wizard, or add a note like "select the tab matching what you chose in Step 2."
Longer-term:
Consider a short GIF/video of gh aw add-wizard running end-to-end so new users know what to expect from the interactive prompts before running it themselves.
For images below the fold (like the sample report screenshot), consider loading="eager" or a fixed placeholder height so first-time readers don't momentarily wonder if content is broken while scrolling.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Summary
Critical Issues Found
None found that block getting started. All three pages loaded correctly, no 404s, and the documented CLI commands (
gh extension install github/gh-aw,gh aw add-wizard,gh aw compile,gh aw run) are shown with clear syntax.One near-miss worth flagging: the sample-output image on the Quick Start page ("The report will look something like this:") initially reported
naturalWidth/Height = 0via script inspection before it scrolled into view — this turned out to be lazy-loading behavior, not an actual broken image (confirmed 200 response and full load after scroll), but a first-time reader skimming quickly could briefly wonder if the image is broken.Confusing Areas
gh aw add-wizardwalks through 6 sub-steps (check prerequisites, select engine, configure auth, add workflow, add skills, trigger run) listed in prose — a numbered list or short animated GIF of the actual interactive prompts would make expectations clearer before running the command for real.What Worked Well
ghCLI v2.0.0+, login, OS) — no assumed context.curl ... install-gh-aw.sh | bash) is provided for users hittinggh extension installauth issues — good anticipation of a common failure mode.gh extension install github/gh-aw@v0.1.0) are clearly explained for teams wanting reproducibility.Recommendations
Quick wins:
Longer-term:
gh aw add-wizardrunning end-to-end so new users know what to expect from the interactive prompts before running it themselves.loading="eager"or a fixed placeholder height so first-time readers don't momentarily wonder if content is broken while scrolling.Screenshots
📎 home.png — Home page, nav and CTAs — asset URL: https://github.com/github/gh-aw/blob/assets/Documentation-Noob-Tester/2d9d7391d69db7d355855c8a464dd0690de5e0df777c9e1d19a86e11d121d4aa.png?raw=true
📎 quick-start-top.png — Quick Start guide, prerequisites and Step 1 — asset URL: https://github.com/github/gh-aw/blob/assets/Documentation-Noob-Tester/138d6349b8dd745cb578d4203f3957c7163135df10e7716fdcac847b369c839f.png?raw=true
📎 quick-start-report-image.png — Quick Start guide, sample "Daily Repo Status Report" output image — asset URL: https://github.com/github/gh-aw/blob/assets/Documentation-Noob-Tester/df8367bcc7d2838e5cef6b52ef49997cab09b5977910d84bcd086ca4cdbab923.png?raw=true
📎 cli-page.png — CLI Commands page, "Day-one commands" table — asset URL: https://github.com/github/gh-aw/blob/assets/Documentation-Noob-Tester/341d1c446c2b370839c896e928271d2cea5a6ca84f26a9cb53452f4d75d24251.png?raw=true
Warning
Firewall blocked 5 domains
The following domains were blocked by the firewall during workflow execution:
accounts.google.comandroid.clients.google.comclients2.google.comwww.google.comwww.gstatic.comTo allow these domains, add them to the
network.allowedlist in your workflow frontmatter:See Network Configuration for more information.
All reactions