@micropage-sh/mcp
Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.
Needs to start: nothing
Returns a result when called with no setup: 0 of 2 tools tried
Contacts while installing and starting: nothing
What it contacted
Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.
Starting it
npx -y @micropage-sh/mcp@0.1.15
Installing a package runs its code. Use a sandbox for anything you have not read.
Calling its tools
2 tools that take no input and look read-only were called once with nothing. 0 returned a result.
whoami | No usable answer replied, but asks for setup: Not signed in. Not logged in to micropage. Run `micropage login` in a terminal, then retry. { "logged_in": false, "auth_mode": |
list_projects | No usable answer tool reported an error: Not logged in to micropage. Run `micropage login` in a terminal, then retry. |
During the calls it contacted nothing it had not already contacted.
Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.
What its tools tell an agent
In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.
Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.
Tools
whoami
Report which micropage account this server acts as: email, plan tier, subscription, and whether it runs on the user's `micropage login` session or on a project deploy token (and if so, which project it is pinned to). Call it first when a micropage tool fails with a login, session or plan error, or when the user asks which account is connected. It never fails for a missing login: it returns logged_in false and a note saying what the user must do (usually run `micropage login` in a terminal). It does not log in, log out, or change anything. Do not call it before every task; the other tools report login and plan problems themselves.
list_projects
List every micropage project in the signed-in account: uuid, numeric id, name, micropage domain, custom domain, live URL, editor URL, and the status and number of each project's active build. Use it to find the `project` value other tools need (uuid, id or domain all work), or when the user refers to a site by name and you do not know which project it is. The active build is the one the editor and CLI work on; it can be a draft newer than what is live, so call get_project for what is actually deployed, failure reasons and custom-domain status. Read-only. Not available with a deploy token, which is limited to its one project.
get_project project
Show one micropage project in detail: live URL, editor URL, the active build (the default publish target), the latest build, the build that is live now, the failure reason if a build failed, and read-only custom-domain status (Cloudflare hostname and certificate state, and the CNAME target).
Use it to check what is live versus drafted before publishing, to explain why a build failed, or to answer "is my custom domain working?". Accepts the project's uuid, numeric id, micropage domain ("acme" or "acme.micropage.live") or custom domain.
It does not return the page source (use get_page_source) or follow a deploy in progress (use get_deploy_status), and it cannot change the custom domain; that is done in the micropage editor.get_page_source project, version
Return the raw .page markup of a micropage project's active build (or of a given build number), plus its stored llms.txt and the filenames uploaded to the project. Call it before editing an existing site: change the returned source and send the whole file back with save_page as landing.page. Pass `version` to read an older build, e.g. to restore it. The source was merged from all .page files when saved, so it comes back as one file. Read-only; it does not show what is live (get_project tells you which build is deployed) and does not download images.
create_project name, domain
Create a new, empty micropage project in the signed-in account and return it together with a complete starter .page file (one of the bundled examples) to adapt. Use it when the user wants a new site. Each call creates another project and counts against the plan's project limit (Pro: 5, Pro+: 20), so check list_projects first when the user may mean an existing site. Nothing is published: write the page, call save_page, and publish only on request. It does not write local files (it is not `micropage projects create`) and cannot set a custom domain; that is done in the editor.
15 more tools
save_page project, pages, llms_txt
Compile .page markup and save it as the project's draft build. It never goes live until publish_build deploys it: call publish_build only when the user asks. Post changes rebuild the live site, not this draft.
Pass the whole site source each time (pages: [{ name: "landing.page", content }], extra .page files are merged after it by name). If the project's active build is a draft or a failed build it is overwritten; otherwise a new draft is created and made active. Returns the build id and number, compiler issues (such as a missing image file) and the editor URL where the user can preview the draft; drafts have no public URL.
Images are not uploaded here: call upload_asset first, then reference the file with `img: <- filename`. Call get_markup_reference before writing markup you are unsure of. Saving also registers the source's forms on the project.publish_build project, build, confirm
Deploy a micropage build to the project's public URL. This changes what visitors see, so only call it when the user explicitly asks to publish, and pass confirm: true. Defaults to the active build (the draft save_page just wrote); pass `build` with the id save_page returned to be exact, or an older build number to roll back to it (that build becomes active). Refused while another build of the project is still publishing. It returns at once with status publishing, after_event_id and an ETA (builds wait about 3 minutes in the queue unless the account is Pro+). Then call get_deploy_status with that build and after_event_id to follow it to the end. It does not save source; call save_page first.
get_deploy_status project, build, after_event_id, wait_seconds
Report the progress of a publish: the build status, the deploy events since after_event_id (queued, built, uploaded, domain wiring, completed or failed), whether it is done, and the live URL. Call it after publish_build or publish_post, passing the build and after_event_id they returned, and keep calling it with the returned after_event_id until done is true. With after_event_id, only a deploy event after it ends the wait: a site rebuild after a post change is queued (up to about 3 minutes unless the account is Pro+) and reports waiting_for_start until it begins. It waits up to wait_seconds (default 20, max 45) per call, sending progress notifications meanwhile. possibly_stuck flags a publish with no events for far longer than a deploy takes. When a newer build of the project takes the deploy over (build.superseded), it follows that build and returns it as `build`; keep calling with that build from then on. When it reports success: Right after deployment.completed the URL can still serve the previous version for up to a minute while Cloudflare's edge caches refresh, and page paths redirect (308) to a trailing slash, so fetch with redirects followed (curl -L). Read-only; it never starts or cancels a deploy.
list_builds project, limit
List a micropage project's builds, newest first: number, id, status (draft, publishing, deployed, failed), created and updated times, failure reason, and which build is active. Use it to find a build to roll back to with publish_build, to see why a publish failed, or to check whether a publish is still running. Each save_page that cannot overwrite a draft adds a build; publishing reuses the build. Read-only. To read a build's source use get_page_source with its number; for what is live right now use get_project.
upload_asset project, filename, source
Upload one image (.png, .jpg, .jpeg, .gif, .webp, .ico, .svg; max 10 MB) to a micropage project's file storage, from a local file path, a public https URL, or base64 bytes. Returns the stored filename, the markup to reference it (`img: <- hero.webp`) and its URL. Use it BEFORE save_page or upsert_post reference the image: page markup like `img: <- hero.webp` and a post's `hero: hero.webp` only resolve to files already stored in the project, and a missing name renders no image. Upload each asset once; calling again with identical content is a no-op (deduped: true, same SHA-256 as the CLI uses). Uploading different content under an existing name deletes the old file and stores the new one. A page that is already live keeps pointing at the old file until it is saved and published again, so re-run save_page and publish_build after replacing an image on a live site. Prefer a new filename to avoid that. Do not use it for non-image files (unsupported), for images already hosted elsewhere that markup can reference by full https URL, or to look up existing files (use list_files). Fails with the plan's storage limit when the project is full.
list_files project
List the files stored in a micropage project (images uploaded by upload_asset, the editor or the CLI), newest first: filename, type, size, SHA-256 and upload date, plus storage used and the plan's storage limit. Use it to see which filenames .page markup can reference with `img: <- name`, to check whether an image is already uploaded, or to check remaining storage before uploading large images. It does not return file URLs (use get_file_url for one) or upload anything (use upload_asset). Read-only.
get_file_url project, filename
Get the URL a stored project file is served from, by its filename (as list_files reports it). Use it when the user wants a link to an uploaded image, or a full URL is needed somewhere a bare filename does not work (for example a post's hero image or an external page). Inside .page markup, prefer the bare filename (`img: <- hero.webp`), which survives re-uploads. It does not upload or list files (use upload_asset / list_files). Read-only.
list_forms project
List the forms on a micropage project's site (registered when the page markup is saved): form uuid, name, page, whether it is the site footer form, and how many submissions each has, not counting spam. Use it to answer "how many signups/messages did my site get?", to find a form's uuid or name before reading its submissions, or to check that a form added in the markup was registered after save_page. It returns counts only, never what visitors typed (that is list_submissions, which the user must enable). Read-only.
list_posts project, slug
List a project's posts (blog posts and newsletter emails), or read one in full. Without `slug`: every post, newest first, with slug, title, visibility, whether it is published and when, whether publishing emails it, and the newsletter list name. With `slug`: that post in full, including body_markdown and every field under the same names upsert_post takes (title, description, visibility, hero, email, list, subject, preview). Call it with a slug before editing a post: upsert_post replaces every field, so start from the current values, and pass the post's revision back to upsert_post. It does not say who will receive an email; preview_post_send does that. Read-only.
upsert_post project, title, slug, body_markdown, description, visibility, hero, email, list, subject, preview, date, revision, confirm_live_update
Create or update one post (a blog post under /content and/or a newsletter email), keyed by slug: an existing slug is updated, a new one is created as a draft. Same as `micropage posts push` for one file; the arguments are the post front-matter fields plus body_markdown (see the micropage://posts/format resource). Every call replaces the whole post: an omitted optional field (description, hero, subject, preview, email/list) is cleared. The exception is `date` (the post's public date, past or today): omitted, the post keeps its current date. To edit, read the post with list_posts(slug) first and pass everything back, with its `revision`: if the post changed since that read (in the editor, the CLI or another session), the save is refused with REVISION_CONFLICT and nothing is written; re-read it, merge, and save again. Saving never sends email and never publishes a draft. But if the post is already published, the save changes the live page at once and rebuilds the site, so it needs confirm_live_update: true after the user agrees. The site rebuild redeploys the live build, so an unpublished page draft left by save_page or the editor stays a draft. It returns build_id and after_event_id to follow the rebuild with get_deploy_status. When every field already matches the saved post (a date that keeps the current one counts as matching), nothing is written, no rebuild is queued, no confirmation is needed, and action is "unchanged". Images are not uploaded here. For the hero pass an https URL or the filename of an asset uploaded with upload_asset; inside body_markdown use hosted URLs only (upload_asset, then get_file_url). To publish afterwards use preview_post_send, then publish_post.
preview_post_send project, slug
Show what publish_post would do for one post, and get the confirmation_token publish_post requires. Call it before every publish and show the user the result. Reports whether publishing emails the newsletter list, the list name, the number of active subscribers it would go to, whether publishing is a re-send (the post is already published, or was emailed before and has since been unpublished; either way publishing emails the whole list again), and whether this server is allowed to send email (MICROPAGE_MCP_ALLOW_SEND). The token is tied to the post as it is now: any edit to the post (upsert_post) or to its list invalidates it, and it expires after 15 minutes or when the server restarts. Read-only: nothing is published or sent.
publish_post project, slug, confirmation_token
Publish a post: put it live on the site under /content and, if it has email enabled with a list, email it to every active subscriber on that list. Publishing a post that is already published, or that was emailed before and has since been unpublished, re-sends the email to the whole list. Same as `micropage posts publish <slug>`. Needs the confirmation_token from preview_post_send, taken after the post's last change; show the user that preview and get their go-ahead first. A stale or missing token is refused with nothing changed. Posts that would send email are refused unless the user has set MICROPAGE_MCP_ALLOW_SEND=1 in this server's config; there is no web-only override, so to publish such a post on the web only, save it with email: false first. When the client supports it the user is also asked to confirm the send directly. The site rebuild redeploys the live build, so an unpublished page draft left by save_page or the editor stays a draft. It returns build_id and after_event_id to follow the rebuild with get_deploy_status. Use unpublish_post to take a post down.
unpublish_post project, slug, confirm
Take a published post off the site: its page 404s and it goes back to draft. The post itself is kept and can be published again later; emails already sent are not recalled. Same as `micropage posts unpublish <slug>`. Needs confirm: true, which you should only pass after the user has agreed. Queues a site rebuild when the project has been published. The site rebuild redeploys the live build, so an unpublished page draft left by save_page or the editor stays a draft. It returns build_id and after_event_id to follow the rebuild with get_deploy_status. To remove a post for good use delete_post instead.
delete_post project, slug, confirm
Delete a post for good, removing it from the site if it was published. There is no undo; to only take it off the site use unpublish_post. Same as `micropage posts rm <slug>`. Needs confirm: true, which you should only pass after the user has agreed. Deleting a slug that does not exist is not an error (deleted: false). Queues a site rebuild when a post was removed and the project has been published. The site rebuild redeploys the live build, so an unpublished page draft left by save_page or the editor stays a draft. It returns build_id and after_event_id to follow the rebuild with get_deploy_status.
get_markup_reference topic, example
Return micropage reference text: the .page grammar, the full reference, the posts format, the agent editing guide, the example index, or one complete example file.
Call this before writing or editing .page markup (and before drafting a post), unless you already have the grammar in context. The vocabulary is closed: tags that are not in the grammar are not rendered, so never guess. Start with topic "grammar"; use "examples_index" then "example" to copy a layout close to what the user wants.
This mirrors the micropage://grammar, micropage://grammar/full, micropage://posts/format, micropage://agent-guide and micropage://examples/{name} resources, for hosts that do not read resources. Local and read-only: no network call, no account needed.Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.
Install facts
| License | MIT |
|---|---|
| Needs | node >=20 |
| Source | github.com/micropage-sh/mcp |
| Install scripts | None declared by it or anything installed with it. |
| Tarball | sha512-zN1T6b4xyiqufJTvQBRd0YRriCTjLqzviteLnb+3JMrt3x8C1PpX6LDYXj3SIZjVljx+wKBoaR32EicD3WU6qQ== |
Check this report yourself
This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.
Report as JSONExact text and tokenHow to verify it
Show it on your README
[](https://bragproof.smaverk.com/r/%40micropage-sh/mcp)
New version out? Ask for a fresh report.