Edgeone Makers Deploy

EdgeOne Makers Deployment Skill

Deploy any project to EdgeOne Makers.

⛔ Critical Rules (never skip)

1. CLI version ≥ 1.6.0 — reinstall if lower. Versions below 1.6.0 lack the non-interactive fixes (whoami fail-fast, --json output) and hang in Agent/CI environments. Never proceed with an outdated version. Anonymous deploy and claim additionally require 1.6.29 — that is a higher, feature-specific floor; see Anonymous Deploy. Do not block a normal authenticated deploy because the CLI is below 1.6.29.

2. Never truncate the deploy URL — this applies to EVERY mention — EDGEONE_DEPLOY_URL includes query parameters (?eo_token=...&eo_time=...) required for access. Without them the page returns 401. Always output the complete URL with full query string. This rule applies to: the primary display, summary tables, footnotes, comparisons, code blocks, present_files calls — every single occurrence of the URL in your reply. Truncation is any removal of the ? and everything after it.

   ❌ WRONG (truncated — will 401):

   Copy code
   https://my-app-w9t0lxe8.edgeone.cool
   

   ✅ CORRECT (full URL):

   Copy code
   https://my-app-w9t0lxe8.edgeone.cool?eo_token=abc123&eo_time=1234567890
   

   Self-check after writing your reply: scan for every instance of the .edgeone.cool domain. Does each one include ?eo_token=? If any doesn't, fix it NOW — the user will get a 401. 3a. Prefer --json when running non-interactively — in Agent/CI/headless contexts, always pass --json to deploy so the result is a single machine-readable line; no need to scrape colored/\r-animated stdout. See Parse Deploy Output. 3b. Use edgeone whoami to check login status — on CLI ≥ 1.6.0, whoami fails fast (exit 1) when not logged in instead of hanging. If it exits 0, the user is already logged in and -t is not needed. Do NOT check cat .edgeone/.token — CLI stores credentials in ~/.edgeone/<hash> files, not a fixed .token path.

3. The deploy URL MUST be placed at the very top of the visible reply body (own line, code block, or heading — not inline, not mid-reply), AND ALSO pinned via present_files (or the IDE's preview tool) to the side panel. Two UI failures to survive: thinking / reasoning / "深度思考" blocks hide content by default; long replies get auto-folded by IDE chat cards, burying anything placed mid-reply. present_files is unaffected by chat folding — that's the second, always-visible channel. Complete URL, no truncation (Rule 2). Example format:

   Copy code
   🌐 Live URL: https://my-project-abc123.edgeone.cool?<auth_query_params>
   

   Then append any other notes (console URL, caveats, etc.).

   Self-check before ending the turn (MANDATORY) — read back what the user will actually SEE (NOT your thinking / reasoning content). Two questions: (a) Is the complete .edgeone.cool URL present at the top of the visible reply, in a code block or heading? (b) Was present_files called with that URL? If either answer is no, send an additional short message containing ONLY the 🌐 Live URL: <full URL> block and call present_files. Do not end the turn until both channels carry the URL. "I already mentioned it in my reasoning" is NOT a substitute for placing it in the visible body.

4. Ask the user to choose China or Global site before browser login. Never assume. (Token login via edgeone login --token auto-detects site, no need to ask.)

5. Prefer Browser Login; fall back to Token only after browser login is confirmed to fail (see Login section for the ~60s fallback threshold and the Agent-in-IDE clarification — WorkBuddy is NOT headless). Token-first only when the user explicitly requests it.

6. After token login, ask if the user wants to save the token locally for future use.

7. Before triggering any browser popup (login / registration), explain the reason and the benefits to the user first — never silently launch a browser window.

8. On any CLI failure, surface the actual error text to the user before retrying, switching commands, or proposing a workaround. Do NOT paraphrase (e.g. don't rewrite Makers project exceeds 40 limit into "maybe a name conflict or permission"). Do NOT silently pivot from makers dev to makers deploy (or vice versa) hoping to bypass — a systemic failure (auth / quota / permission) hits both with the same cause. Quote the raw error, name the root cause, then propose the fix or ask the user.

9. Write every user-facing message in the user's own language — all prose, blockquotes, and prompts in this skill are written in English purely to specify meaning; none of it is a string to paste. If the user writes to you in Chinese, speak Chinese — the login explanation, the site choice, the token question, the deploy result, all of it. Likewise for any other language. Emitting this skill's English strings into a non-English conversation is a bug. Only literal CLI commands, flags, env var names, and JSON field names stay verbatim. ⚠️ Where this skill gives a fixed template (Steps 2 and 4 of Anonymous Deploy), translating it is required but rewriting or extending it is not allowed — match it line for line.

Rules 11-14 apply to the anonymous deploy / claim flow only:

11. The claim command's parameter is --sid, NOT --token — edgeone makers claim --sid <anonymous-token>. The -t / --token flag on claim is the account API token, an entirely different credential. Passing the anonymous token to -t fails. Product documentation showing claim --token <anonymous-token> is wrong; trust this rule.
12. Present the anonymous deploy result with the fixed template in Step 4 — it is a template, not an example: reproduce it exactly, substituting only the URL and the claim link. Never show the raw expiresAt timestamp, never state a different duration, and never add lines to it. 60 minutes is the product's stated claim window and the conservative instruction, so that is what the template says even though the token's observed lifetime can be longer. It frames the link as expiring, not the project as being deleted.
13. The claim link you present must come from edgeone login --claim, never the bare claimUrl — right after an anonymous deploy, run edgeone login --claim --local --json in the background and put its claimLoginUrl in the template. Only that link carries the login → claim → callback chain that lets the CLI learn the login state and auto-link the project; the bare claimUrl from deploy output claims without telling the CLI, so the user's next deploy would create a second anonymous project. Never print claimCommand, edgeone makers claim --sid ..., or any other command as the user's way to claim — in sandboxed IDEs like WorkBuddy the user has no terminal and literally cannot run it. makers claim is only a fallback for headless/CI environments, run by you, never displayed.
14. Keep the claim pitch minimal — do not over-promise, and do not teach domains — say only that signing in keeps the project. ❌ Never write "permanently yours", "no time limit or access restrictions", "unlimited", or anything implying the URL then works unconditionally forever: a claimed project may still need a custom domain, and mainland-China access can require ICP filing, so those claims are false. ❌ Also do not volunteer custom domains, ICP filing, DNS, or console navigation while the user is just deciding whether to claim — that front-loads complexity onto someone who only wanted a live URL. The claim page owns the follow-up flow. Answer such topics only when the user asks.

Environment Setup

Before executing any edgeone CLI command (install, login, deploy, etc.), set the following environment variable in the current shell session:

bashCopy code
export PAGES_SOURCE=skills

Or prefix each command inline:

bashCopy code
PAGES_SOURCE=skills edgeone makers deploy

This tells the platform that the deployment is triggered from an AI skill context.

Deployment Flow

Run these checks first, then follow the decision table. Run them silently — do not narrate versions, login-state probes, or eligibility results to the user (e.g. never say "CLI 版本 x.x（支持匿名部署）" or "纯静态、无依赖、可直接发布"). Speak only when a check changes the outcome (needs upgrade, needs login).

bashCopy code
# Check 0: Set environment variable (required before any edgeone command)
export PAGES_SOURCE=skills

# Check 1: CLI installed and correct version? (must be >= 1.6.0; anonymous deploy needs >= 1.6.29)
edgeone -v

# Check 2: Already logged in? (whoami fails fast, won't hang)
edgeone whoami
# If exit 0 → logged in, no -t needed
# If exit 1 → not logged in, need token or browser login

# NOTE: This auth gate is for `deploy` (account-bound upload to your EdgeOne
# account). For `edgeone makers dev` (local preview), login is ONLY required
# when the project uses Blob/credentialed backends — a pure-static dev needs
# no login. When Blob IS used, the chain is: Blob → must be linked → linking
# requires login, so login before linking/starting dev. See makers-storage /
# makers-env-adaption for the dev auth rule and the link chain.

# Check 3: Project already linked?
cat edgeone.json 2>/dev/null

Decision Table

Install CLI

bashCopy code
npm install -g edgeone@latest

Verify: edgeone -v — confirm output is 1.6.0 or higher. Retry installation if not. (Versions < 1.6.0 hang on whoami/login in non-interactive environments and lack --json.) Anonymous deploy and claim need 1.6.29; if latest is still below that, those two features are simply unavailable — normal authenticated deploy works fine.

Login

0. Explain the registration/login step

Before triggering any login flow, explain to the user why this step is needed and what to expect. Do not silently launch a browser window.

Tell the user:

You need to log in or register an EdgeOne Makers account. Here's what to expect:

• Why login is required: Deployment uploads your build output to your own account, generating a unique access URL and project record.
• What you get for free: EdgeOne Makers offers a free tier with global CDN acceleration, automatic HTTPS, and custom domain binding — typically more than enough for personal projects.
• What happens next: I'll run edgeone login, and your default browser will open the Tencent Cloud login page. Please complete the login/registration and authorize access, then come back here.
• If you get stuck: If the browser doesn't open, or the CLI keeps waiting after you've logged in, let me know — I'll switch to Token login instead.

If the user does not respond within ~60 seconds (no browser popup or no progress reported), proactively ask about their status (whether the browser opened, any errors, or if they want to switch to Token login). Do not wait indefinitely.

1. Ask the user to choose a site, then ALWAYS pass --site

Use the IDE's selection control (ask_followup_question) before running any login command:

Choose your EdgeOne Makers site:

• China — For users in mainland China (console.cloud.tencent.com)
• Global — For users outside China (console.intl.cloud.tencent.com)

⚠️ CRITICAL: After the user chooses, you MUST invoke login with an explicit --site <china|global> flag (e.g. edgeone login --site china). NEVER run a bare edgeone login (without --site) when driven by an Agent / skill. A bare login in a non-interactive context fails fast asking for --site (it no longer pops an interactive site-picker that would hang). The site choice is meant to happen here in the conversation, not inside the CLI.

2. Login methods reference

Two login methods are available. Per Rule 6, always try Browser Login first; the table below is a reference for when each method applies, not a decision procedure — do not use it to guess the environment.

Browser Login

bashCopy code
# China site
edgeone login --site china

# Global site
edgeone login --site global

Wait for the user to complete browser auth. The CLI prints a success message when done.

⚠️ Browser Session Reuse Trap: If the user previously logged into a different site (e.g., logged into Global site before, now trying China site, or vice versa), the browser may silently reuse the old Tencent Cloud session. The CLI will appear to succeed, but actually binds to the wrong account — subsequent deploy will fail with auth errors or whoami shows an unexpected account.

If this happens, guide the user to:

1. Click "Sign in with a different account" on the login page; or
2. Log out from all Tencent Cloud consoles (both console.cloud.tencent.com and console.intl.cloud.tencent.com) first, then re-run edgeone login.

Token Login

Two methods available:

Method A: edgeone login --token (persistent, recommended)

bashCopy code
edgeone login --token <token>

Auto-detects china/global from the token — no --site flag needed. Persists login state for subsequent commands.

💡 Reuse the token from a prior browser login — no console trip needed. edgeone login --site <x> (browser) auto-generates an API Token and writes it to ~/.edgeone/<hash> (JSON with value.Token). You can reuse that Token value directly as EDGEONE_PAGES_API_TOKEN="<Token>" or -t <Token> for makers dev/deploy in headless/agent contexts, instead of creating a new token in the console.

Method B: Pass -t directly in deploy (per-invocation)

Token is used for that single deploy only; no persistent login state is saved.

bashCopy code
edgeone makers deploy -t <token>

⚠️ Important: edgeone whoami does NOT support a -t flag. Do NOT attempt to verify a token with whoami -t <token>. When the user provides a token, skip login checks entirely and go straight to deploy.

Guide the user to obtain a token:

1. Go to the console:
   • China: https://console.cloud.tencent.com/edgeone/pages?tab=settings
   • Global: https://console.intl.cloud.tencent.com/edgeone/pages?tab=settings
2. Find API Token → Create Token → Copy it

⚠️ Remind the user: the token has account-level permissions. Never commit it to a repository.

3. Offer to save the token locally

After the user provides a token, ask:

Save this token locally for future deployments?

• Yes — Save to .edgeone/.token (auto-used next time)
• No — Use for this deployment only

If Yes:

bashCopy code
mkdir -p .edgeone
echo "<token>" > .edgeone/.token
grep -q '.edgeone/.token' .gitignore 2>/dev/null || echo '.edgeone/.token' >> .gitignore

Confirm to the user: "✅ Token saved to .edgeone/.token and added to .gitignore."

Deploy

Browser-authenticated deploy (Makers projects)

bashCopy code
# Project already linked (edgeone.json exists)
edgeone makers deploy

# New project (no edgeone.json)
edgeone makers deploy -n <project-name>

<project-name>: auto-generate from the project directory name. The first deploy creates edgeone.json automatically.

Token-based deploy (Makers projects)

First check for a saved token:

bashCopy code
cat .edgeone/.token 2>/dev/null

• Saved token found → use it, tell the user: "Using saved token from .edgeone/.token"
• No saved token → ask the user to provide one (see Token Login above)

bashCopy code
# Project already linked
edgeone makers deploy -t <token>

# New project
edgeone makers deploy -n <project-name> -t <token>

The token already contains site info — no --site flag needed.

After a successful deploy with a manually-entered token, ask if the user wants to save it (see "Offer to save the token locally" above).

Deploy to preview environment

bashCopy code
edgeone makers deploy -e preview

Non-interactive / Agent / CI deploy (recommended: --json)

When running inside an Agent, CI, or any non-TTY context, add --json so the final result is emitted as a single machine-readable line — no scraping of colored stdout:

bashCopy code
edgeone makers deploy -n <project-name> --json
edgeone makers deploy -n <project-name> -t <token> --json

Makers Agent Projects deploy

For projects with agents/ directory (AI Agent projects), use edgeone makers deploy which auto-runs build:

bashCopy code
edgeone makers deploy -n <name> -t <token> --json
edgeone makers deploy -n <name> -t <token> --json -e preview

Note: edgeone makers deploy automatically runs build before deploying — no separate edgeone makers build step needed.

Build behavior

The CLI auto-detects the framework, runs the build, and uploads the output directory. No manual config needed.

Anonymous Deploy (no login required)

When the user is not logged in and has no token, they can deploy anonymously and claim the project later. Requires CLI >= 1.6.29. On 1.6.0–1.6.28 this feature does not exist — use Login or a token instead.

Step 1: Exclusion check — run this FIRST

Anonymous deploy builds with an empty environment: it skips remote env-var pull and AI-gateway credential injection, because both need authentication. Projects depending on either will deploy successfully but break at runtime.

bashCopy code
# Is this an Agent project?
ls agents/ 2>/dev/null && echo "AGENT_PROJECT"

# Does it use Blob storage? (checks source files and the dependency declaration)
grep -rEl "@edgeone/pages-blob" \
  --include="*.ts" --include="*.js" --include="*.mjs" --include="*.cjs" \
  --include="*.tsx" --include="*.jsx" --include="*.vue" --include="*.svelte" \
  . 2>/dev/null | head -1
grep -l '"@edgeone/pages-blob"' package.json 2>/dev/null

Non-empty output from either grep means the project uses Blob.

⛔ All checks in this flow are silent — never narrate them to the user. Do not report things like "CLI 版本 x.x（支持匿名部署）", "页面是纯静态、无 Blob/KV 依赖，可直接发布", or any other version/eligibility check result. These are internal reasoning; a non-technical user cannot act on them and should never see them. The only time you speak about a check is when it changes the outcome — e.g. the project needs login (below), or the CLI is too old and needs upgrading. Passing checks produce no message at all.

KV cannot be detected this way — you must ask. A KV namespace is bound in the console and exposed as a global variable whose name the user chose (e.g. my_kv), so there is no package import to grep for. There is no @edgeone/pages-kv package. Ask the user directly:

Does this project use KV storage?

If either check hits, or the user says the project uses KV, do not deploy anonymously. Go to Login and tell the user why (note the user-facing wording: "without login", never "anonymous"):

This project needs environment variables / AI gateway credentials or a storage binding, which a login-free deploy cannot provide. The site would load but those features would fail. Let's log in so it works properly.

Plain static sites and frontend-framework projects with no such dependency may proceed.

If the user acknowledges the limitation and still wants to publish without logging in, you may proceed — but state prominently in your result that AI and storage features will not work until the project is claimed and configured (again: phrase it as "login-free / 免登录", never "anonymous / 匿名").

Step 2: Decide the path — default to anonymous only when the task fits, otherwise ask

Go anonymous directly, without asking, only when the task itself is anonymous-native — disposable or unattended by nature. Clear signals:

• Scheduled / automated jobs — e.g. "每日定时生成一个页面并部署", cron pipelines, any task that must run with nobody watching
• Disposable one-offs — a throwaway preview, a quick demo, "先看看效果", anything the user frames as temporary or just-for-now
• Explicit anonymous intent — the user says "免登录", "不用登录", "anonymous", or "don't make me sign up"

In these cases the 60-minute expiry + claim-later model is the right answer, so asking is pure friction. Go straight to Step 3, and present the result with the fixed template in Step 4 (claim link + 60-minute window) — the template already tells the user how to keep it if they change their mind.

Everything else — ask. If the task isn't obviously disposable — anything the user might want to keep, share widely, or build on — present the choice. Do NOT try to judge the environment to skip the question: sandbox/TTY signals are unreliable and guessing wrong either needlessly interrupts a human or silently deploys something they wanted to keep.

If the environment genuinely gives you no way to ask at all (no TTY, no question tool), asking is impossible — that's a constraint, not a judgment call. Deploy with --anonymous --json and make the claim link and the 60-minute window unmissable in your result.

How to ask (when asking): present the choice with the IDE's selection control (e.g. ask_followup_question). The option labels deliberately avoid the word "anonymous" — it is jargon and confuses non-technical users. Present these two options exactly, in the user's language. Do not paraphrase the labels, do not add caveats to the options themselves, and do not mention "anonymous" to the user at all — when you must name the concept, call it "login-free deployment" / 「免登录部署」, never "anonymous deploy / 匿名部署". The option text must be placed in the label exactly as is; it must not be split into the description.

English-speaking user — present exactly these two options:

• Publish now, log in later — No login needed. Get a preview link right away; you can log in and claim the site whenever you're ready.
• Log in, then publish — Log in to your EdgeOne Makers account first; the site is saved straight to your account.

Chinese-speaking user — present exactly these two options:

• 直接发布，无需登录
• 登录后发布

Map their pick to the flow:

• Publish now, log in later → this is the anonymous deploy path — continue to Step 3.
• Log in, then publish → go to Login (browser flow), then continue with the normal deploy.

⛔ Do not over-promise what claiming gives them. Say the project is kept / saved to their account, and nothing more. Specifically, do not say "no access restrictions", "permanently yours", or anything implying the URL is then unconditionally public and final — a claimed project can still need a custom domain and, for mainland-China access, ICP filing. Do not raise custom domains, ICP filing, or DNS at this point either: the user is deciding whether to log in, and those concepts are noise here. The claim page walks them through next steps.

Step 3: Deploy

bashCopy code
export PAGES_SOURCE=skills
edgeone makers deploy --anonymous --json

Add --site china or --site global only when the site must be pinned (the user told you which, or CI needs determinism). Otherwise omit it — the CLI detects the site itself from its own egress IP. Do not probe for the site yourself and do not synthesise --site from your own check.

Do not pass -n, -e, or --area — they are ignored under --anonymous. The project name is generated automatically.

Step 4: Start the claim listener, then present the result

Parse the last line of the deploy stdout as JSON (human-readable output precedes it). Fields are camelCase: url, projectId, deploymentId, anonymousToken, claimUrl, claimCommand, expiresAt, site.

Immediately start the claim listener in the background — it must be listening before the user clicks the claim link, or the login state never reaches the CLI and the next deploy would create a second anonymous project:

bashCopy code
edgeone login --claim --local --json

Run it with run_in_background. It prints a {"status":"waiting","claimLoginUrl":...} JSON line right away, then keeps listening (up to 60 minutes). Grab claimLoginUrl from that line. ⛔ This URL — not the claimUrl field from the deploy JSON — is the claim link you show the user. Only this link chains login → claim page → callback to the CLI. The bare claimUrl from deploy output would claim the project without telling the CLI, and the user's next deploy would go to a new anonymous project instead of the claimed one.

⛔ The listener is invisible to the user — never narrate it. Do not tell the user a listener/监听器 is running, that you are "waiting for the claim callback", or any other internal mechanism status (state files, JSON parsing, ports). The user-facing surface is exactly one thing: the template below. Claiming is presented as "open the link and sign in", not as a process you are watching.

⛔ Send the template as its own message, after the deploy command has fully returned. Do not emit it in the same message in which you ran the deploy, do not attach it to the command's output, and do not print it while the command is still running. In WorkBuddy, content emitted alongside a running/finished command can be folded into a collapsible process message — if that happens, the user never sees the URL or the claim link. Wait for the Bash call to end, then send the template alone as your next reply.

⛔ This is a fixed template, not a suggestion. Reproduce it exactly: same four lines, same order, same emoji, same separator. Substitute only <url> (from the deploy JSON) and <claimLoginUrl> (from the login --claim output). Do not reword, merge, split, or reorder lines. Do not add a sentence before or after it. Do not append project ID, deployment ID, console URL, expiresAt, next steps, or commentary of any kind.

English-speaking user — emit exactly this:

🌐 Live URL: <url>

---

⏳ This link expires in 60 minutes.

👉 Claim this project — please claim it within 60 minutes.

Chinese-speaking user — emit exactly this:

🎉 部署成功 🌐 访问地址：<url>

---

⏳ 该链接 60 分钟后失效。

👉 认领这个项目 —— 请在 60 分钟内完成认领。

For any other language, translate this template and nothing more — keep the four lines, the emoji, the separator, and the exact same content. Do not take the freedom to add or explain.

Afterwards, stop. Anything you feel like adding here — what claiming unlocks, custom domains, ICP filing, DNS, console navigation, an offer to claim on their behalf — is prohibited by critical rules 13 and 14. Answer those topics only if the user asks.

After the user claims

The background login --claim process exits on its own when the console relays the result. When it exits 0 with {"status":"success",...,"linked":true}:

• the CLI is logged in (credentials saved, plus .edgeone/auth.json from --local),
• the project is linked (.edgeone/project.json),
• .edgeone/anonymous.json is removed.

Tell the user plainly, in their language: the project is now in their account, and any further deploys will update that same project. Then use the normal deploy flow (edgeone makers deploy --json) for all later deploys.

If the process exits non-zero (the user never completed the link within 60 minutes), the anonymous project is still unclaimed: re-run edgeone login --claim --local --json for a fresh link and present it again in the same template.

If the user asks you to "帮我认领" / "claim it for me": point them to the same link again while the listener is still running; if it already exited, re-run the command for a fresh link. Never print the CLI claim command — edgeone makers claim --sid remains only for headless/CI environments with no browser (see critical rule 13).

For the full JSON schema, rate limits, error codes, state-file fields, and site-resolution rules, see references/anonymous-deploy.md.

⚠️ Parse Deploy Output (Critical)

Preferred: --json

When deploy is run with --json, the last line of stdout is a single JSON object — parse that directly, no regex / ANSI cleanup needed:

jsonCopy code
{"status":"success","url":"https://my-project-abc123.edgeone.cool?<auth_query_params>","type":"preset","projectId":"makers-xxxxxxxx","deploymentId":"dp-xxxx","consoleUrl":"https://console.cloud.tencent.com/edgeone/pages/project/makers-xxxxxxxx/deployment/xxxxxxx"}

On failure the last line is {"status":"error","error":"<message>"} and the process exits non-zero.

Use url (full, with query string), projectId, and consoleUrl directly.

Fallback: text output (no --json)

After edgeone makers deploy succeeds, the CLI outputs:

Copy code
[cli][✔] Deploy Success
EDGEONE_DEPLOY_URL=https://my-project-abc123.edgeone.cool?<auth_query_params>
EDGEONE_DEPLOY_TYPE=preset
EDGEONE_PROJECT_ID=makers-xxxxxxxx
[cli][✔] You can view your deployment in the EdgeOne Makers Console at:
https://console.cloud.tencent.com/edgeone/pages/project/pages-xxxxxxxx/deployment/xxxxxxx

Extraction rules:

Show the user — the deploy URL MUST be at the very top of the visible reply AND pinned via present_files to the side panel (see Rule 4 for why both channels are required):

⚠️ URL Integrity Rules (read before composing your reply):

🌐 Live URL: https://my-project-abc123.edgeone.cool?eo_token=abc123&eo_time=1234567890

---

• Console URL: https://console.cloud.tencent.com/edgeone/pages/project/...

ℹ️ Note: This preview URL is for quick deployment verification. When accessed from mainland China, the link may become restricted (e.g., 401) after some time or when shared, due to domain ICP filing status or CDN acceleration policies. For long-term stable public access, bind a custom domain with proper ICP filing.

Error Handling

For CLI command reference, environment variables, local dev setup, and token management details, see references/command-reference.md.