Skip to content

ViralCat MCP server

com.viral-cat/viral-catviral-cat.comstreamable-httpv1.0.0Scanned

Schedule and publish social posts across platforms, with drafts, media uploads and analytics.

21 tools: 2 can delete or cancel things, 1 do something that can't be undone, 12 change things and 6 only read.

Tools: 2 delete, 1 irreversible, 12 write, 6 read.
  • Reads private data · 0
  • Reads text others write · 21
  • Can send data out · 5
  • 2 findings

Tools listed without sign-in. Remote posture

With Anlyon

2 tools here can delete or cancel things and 1 does something that can't be undone. Govern them.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the ViralCat MCP server can do

The ViralCat MCP server is published by viral-cat.com, a domain the MCP Registry verified. It has 21 tools. It can delete or cancel things (2 tools: cancel_post and update_brand), do things that cannot be undone, such as sending, paying or publishing (post_performance) and create or change data (12 tools: content_queue, campaign_summary, diagnose_failures, create_draft and 8 more). 6 tools only read.

21 tools bring in text other people can write (list_brands, get_brand_details, list_posts and 18 more); 5 tools can send data out of your system (post_performance, schedule_post, retry_failed_posts and 2 more).

The fixed checks matched 2 findings (annotation-conflict and credentials-mention). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in, and its tools changed once since Anlyon Radar first saw them.

Last scanned Nov 14, 2026, version 1.0.0.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

21 tools

Deletes or cancels 2Claude Code: deny
  • cancel_postWRITE (safe): cancel a queued post, or move it to a new slot.Reads text others writeNo findings

    Why it is in this group

    • The name "cancel_post" says "cancel": a word for removing or ending something.
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (safe): cancel a queued post, or move it to a new slot. Omit scheduled_date to CANCEL (status -> cancelled; the publisher ignores it). Pass scheduled_date to RESCHEDULE to that time instead (ISO 8601, at least 15 minutes out; a draft keeps its draft status). Works on draft or scheduled posts (cancel or reschedule); also works on FAILED posts but cancel-only (omit scheduled_date) — this is the way to clear a dead failed-post backlog, since failed has no other exit. Posts already posting/posted/cancelled are immutable. Get post ids from content_queue, list_posts or diagnose_failures.

    Annotations the server declares

    • title: Cancel Post
    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: true
    • openWorldHint: true
  • update_brandWRITE (gated): correct a brand's content profile: description, taglines, hashtags, CTA and website links, content type, and the auto-post switch.Reads text others writeNo findings

    Why it is in this group

    • The name "update_brand" says "update": a word for a change.
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (gated): correct a brand's content profile: description, taglines, hashtags, CTA and website links, content type, and the auto-post switch. This is the repair path for a brand carrying wrong copy, which otherwise makes every caption for it wrong, and the off switch for a brand whose generator is posting against your content rules. Only the fields you pass change; the rest keep their stored values. Cannot rename a brand (the name also names its Drive media folder) and cannot touch posting frequency/day or discount settings. Queues nothing and publishes nothing itself.

    Annotations the server declares

    • title: Update Brand
    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: true
    • openWorldHint: true
Can't be undone 1Claude Code: ask
  • post_performancePublish-outcome metrics over a rolling window (default 30 days): posted/failed/cancelled totals, success rate, per-platform breakdown, and the most recent published posts with perm…Reads text others writeCan send data out1 finding

    Why it is in this group

    • The name "post_performance" says "post": a word for a change.
    • The description's first verb is "publish": a word for a change that another call does not undo.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "post_performance" says "post": it sends or publishes something to other people.

    Findings

    • Warn

      The server marks "post_performance" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "post_performance" says "post": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Publish-outcome metrics over a rolling window (default 30 days): posted/failed/cancelled totals, success rate, per-platform breakdown, and the most recent published posts with permalinks plus their latest stored engagement snapshot (views/likes/comments/shares; available for youtube/tiktok/twitter/threads). Snapshots are stored, not live: call refresh_post_metrics first if the numbers must be current. Optionally scope by brand_id and/or platform.

    Annotations the server declares

    • title: Post Performance
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Changes things 12Claude Code: ask
  • campaign_summaryPer-brand campaign rollup (brands are ViralCat's campaign unit): posted/scheduled/failed/draft counts, platforms used, next scheduled date, last posted date, and auto-post settings…Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Per-brand campaign rollup (brands are ViralCat's campaign unit): posted/scheduled/failed/draft counts, platforms used, next scheduled date, last posted date, and auto-post settings. Optionally scope to one brand_id.

    Annotations the server declares

    • title: Campaign Summary
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • content_queueThe forward content pipeline: drafts awaiting approval first, then scheduled posts in publish order.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    The forward content pipeline: drafts awaiting approval first, then scheduled posts in publish order. Use to review what is queued before it goes out, and to find draft ids for schedule_post.

    Annotations the server declares

    • title: Content Queue
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • create_brandWRITE (gated): create a brand (content profile) so create_draft and schedule_post have something to target.Reads text others writeNo findings

    Why it is in this group

    • The name "create_brand" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (gated): create a brand (content profile) so create_draft and schedule_post have something to target. Every other write here needs a brand_id, so an agent had no way to set one up. Returns the brand_id and the Drive folder the publisher will read its media from (Viral-Cat/_Res/{name}/Vids). Publishes nothing and reaches no audience.

    Annotations the server declares

    • title: Create Brand
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • create_draftWRITE (safe): create a DRAFT post for a brand.Reads text others writeNo findings

    Why it is in this group

    • The name "create_draft" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (safe): create a DRAFT post for a brand. Drafts are never picked up by the publisher, so nothing goes live; a human approves in the dashboard or promotes it later with schedule_post. scheduled_date is a proposed slot (default: 7 days out).

    Annotations the server declares

    • title: Create Draft
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • create_media_uploadWRITE (gated): open a Google Drive upload session for media you hold locally, so a clip rendered on your own machine never has to be published publicly just so ViralCat can fetch i…Reads text others writeNo findings

    Why it is in this group

    • The name "create_media_upload" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (gated): open a Google Drive upload session for media you hold locally, so a clip rendered on your own machine never has to be published publicly just so ViralCat can fetch it back. Returns an uploadUrl - PUT the bytes to it yourself (no auth header) and Drive stores the file; the response Location/id is the fileId for create_draft or schedule_post. When brand is given the file lands in the folder the generator actually reads for that media type: Viral-Cat/_Res/{brand}/Imgs for an image, Vids for a video, Music for audio. Without brand it defaults to the Viral-Cat folder. Video, image or audio up to 100MB. Publishes nothing and reaches no audience.

    Annotations the server declares

    • title: Create Media Upload
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • diagnose_failuresGroup failed posts by platform + normalized failure reason: counts, date range, affected brands, a sample raw error, a recoverable flag (retry_failed_posts can re-queue it) and a c…Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Group failed posts by platform + normalized failure reason: counts, date range, affected brands, a sample raw error, a recoverable flag (retry_failed_posts can re-queue it) and a concrete suggested fix per group. Answers "what is broken and why" in one call. Read-only.

    Annotations the server declares

    • title: Diagnose Failures
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • reconnect_platformWRITE (safe): attempt a server-side token refresh for a platform whose access token has lapsed, updating the stored connection's token/expiry.Reads text others write1 finding

    Why it is in this group

    • The description's first verb is "write": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credentials".

      Matched text

      description: "...ublish or touch content, only refreshes credentials; safe to retry. Never exposes tokens."

      What this check looks for
    Description, as the server wrote it
    WRITE (safe): attempt a server-side token refresh for a platform whose access token has lapsed, updating the stored connection's token/expiry. Works without a browser when a refresh token is on file (twitter/youtube/google_drive self-heal this way); returns needsManualReauth + a dashboard link when the platform genuinely requires a manual browser reconnect (e.g. threads). Cannot publish or touch content, only refreshes credentials; safe to retry. Never exposes tokens.

    Annotations the server declares

    • title: Reconnect Platform
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • refresh_post_metricsWRITE (gated): fetch current engagement (views/likes/comments/shares) for already-published posts and store today's snapshot, then read it back with post_performance.Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "refresh_post_metrics" says "post": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "refresh_post_metrics" says "post": it sends or publishes something to other people.
    Description, as the server wrote it
    WRITE (gated): fetch current engagement (views/likes/comments/shares) for already-published posts and store today's snapshot, then read it back with post_performance. This is the only way to update the numbers without opening the dashboard Stats tab. Supported platforms: youtube, tiktok, twitter, threads; LinkedIn and the rest expose no engagement API. A 60-minute cooldown applies per account and protects X's free-tier read quota, so a call inside it returns skipped:true and changes nothing. Publishes nothing and reaches no audience.

    Annotations the server declares

    • title: Refresh Post Metrics
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • reschedule_postWRITE (gated): move ONE post to a new future slot (ISO 8601, at least 15 minutes out).Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "reschedule_post" says "reschedule": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "reschedule_post" says "reschedule" and "post": it writes something other people read.
    Description, as the server wrote it
    WRITE (gated): move ONE post to a new future slot (ISO 8601, at least 15 minutes out). Unlike cancel_post this also accepts FAILED posts: a failed post is re-queued (status -> scheduled, error cleared). Draft and scheduled posts keep their status. Refuses posts that would structurally fail again (archived brand, incompatible media, Facebook with no Page).

    Annotations the server declares

    • title: Reschedule Post
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • retry_failed_postsWRITE (gated): bulk re-queue recoverable failed posts (default reason classes: token_expired, missed_window, platform_transient_error, drive_media_missing).Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "retry_failed_posts" says "post": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "retry_failed_posts" says "post": it sends or publishes something to other people.
    Description, as the server wrote it
    WRITE (gated): bulk re-queue recoverable failed posts (default reason classes: token_expired, missed_window, platform_transient_error, drive_media_missing). dry_run DEFAULTS TO TRUE and only returns the plan; pass dry_run=false to actually reschedule. Retries are staggered per platform (spacing_minutes apart) and capped at limit, so a bulk retry cannot dogpile a platform API. Posts that would structurally fail again (archived brand, incompatible media, Facebook with no Page) are skipped with the reason. Use diagnose_failures first to see what is recoverable.

    Annotations the server declares

    • title: Retry Failed Posts
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • schedule_postWRITE (gated): queue a post for FUTURE publishing, at least 15 minutes out (immediate posting is not available over MCP).Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "schedule_post" says "schedule": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "schedule_post" says "schedule" and "post": it writes something other people read.
    Description, as the server wrote it
    WRITE (gated): queue a post for FUTURE publishing, at least 15 minutes out (immediate posting is not available over MCP). Either pass post_id to promote an existing draft, or brand_id + platform to create a new scheduled post. The publishing cron sends it at the scheduled time.

    Annotations the server declares

    • title: Schedule Post
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • upload_mediaWRITE (gated): fetch a public media URL and store it in the connected Google Drive, returning the Drive fileId to pass to create_draft or schedule_post.Reads text others writeNo findings

    Why it is in this group

    • The name "upload_media" says "upload": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    WRITE (gated): fetch a public media URL and store it in the connected Google Drive, returning the Drive fileId to pass to create_draft or schedule_post. This uploads to YOUR OWN Drive; it publishes nothing and reaches no audience. Accepts video, image or audio up to 100MB. With brand, files land in the folder the generator actually reads for that media type: Viral-Cat/_Res/{brand}/Imgs for an image, Vids for a video, Music for audio. Without brand it defaults to the Viral-Cat folder, and folder_id overrides both.

    Annotations the server declares

    • title: Upload Media
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
Only reads 6Show 6 read-only tools
  • get_brand_detailsFull content profile for one brand: description, taglines, hashtags, CTA/website URL, niche (category/subcategory), content type, target platforms, and active discount.Reads text others writeNo findings

    Why it is in this group

    • The name "get_brand_details" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "website": it reads web pages, which anyone can write.
    Description, as the server wrote it
    Full content profile for one brand: description, taglines, hashtags, CTA/website URL, niche (category/subcategory), content type, target platforms, and active discount. Use this to write on-brand captions BEFORE calling create_draft or schedule_post. Read-only; returns no secrets.

    Annotations the server declares

    • title: Get Brand Details
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_pipeline_summaryHigh-level snapshot of the posting pipeline: active brand count, a status breakdown (how many scheduled/posted/failed/etc.), a per-platform breakdown, how many posts are due in the…Reads text others writeNo findings

    Why it is in this group

    • The name "get_pipeline_summary" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "post": it reads posts, which other people can write.
    Description, as the server wrote it
    High-level snapshot of the posting pipeline: active brand count, a status breakdown (how many scheduled/posted/failed/etc.), a per-platform breakdown, how many posts are due in the next 7 days, and the next scheduled date. Use for 'how's the pipeline?' style questions.

    Annotations the server declares

    • title: Get Pipeline Summary
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • list_brandsList brands (content profiles), newest first, with per-brand post counts (posted/scheduled) and the next scheduled date.Reads text others writeNo findings

    Why it is in this group

    • The name "list_brands" says "list": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "list" and "post": it reads posts, which other people can write.
    Description, as the server wrote it
    List brands (content profiles), newest first, with per-brand post counts (posted/scheduled) and the next scheduled date. Use to see which brands exist and how active each is. Excludes archived brands. Returns `total` (all active brands) next to `count` (rows returned): when `truncated` is true the roster was cut off by `limit`, so raise it (max 50) before concluding a brand does not exist.

    Annotations the server declares

    • title: List Brands
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • list_connectionsList connected platform accounts (channels) and their health: platform, account username, connected flag, posting-enabled flag, and token expiry.Reads text others writeNo findings

    Why it is in this group

    • The name "list_connections" says "list": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    List connected platform accounts (channels) and their health: platform, account username, connected flag, posting-enabled flag, and token expiry. Use to check which platforms are wired up and whether any connection has expired. Never returns access or refresh tokens.

    Annotations the server declares

    • title: List Connections
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • list_postsList posts across all platforms, newest scheduled first.Reads text others writeNo findings

    Why it is in this group

    • The name "list_posts" says "list": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe name "list_posts" says "list" and "post": it reads posts, which other people can write.
    • untrusted-inputThe description's first sentence says "list" and "post": it reads posts, which other people can write.
    Description, as the server wrote it
    List posts across all platforms, newest scheduled first. Each row is one publication on one platform. Filter by status (draft|scheduled|posting|posted|failed|cancelled) and/or platform (youtube|tiktok|instagram|...). Returns caption, brand name, permalink, error message when present, and publishWarning for posts that succeeded with an anomaly (e.g. published without media).

    Annotations the server declares

    • title: List Posts
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • preflight_checkREAD-ONLY: validate a post BEFORE it is scheduled or published, so a problem is caught up front instead of discovered as a failure afterwards.Reads text others writeNo findings

    Why it is in this group

    • The description's first verb is "read": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "read" and "post": it reads posts, which other people can write.
    Description, as the server wrote it
    READ-ONLY: validate a post BEFORE it is scheduled or published, so a problem is caught up front instead of discovered as a failure afterwards. Checks the exact things that produced the failed-post backlog: incompatible media for the platform (e.g. video to Pinterest), a disconnected or unrefreshable platform token (LinkedIn has no refresh token on this API tier and needs a manual browser reconnect roughly every 60 days — that's an expected state, not a bug), Facebook posting to an account with no Page (impossible via the Graph API), content referencing an already-passed date/season/promo code, and X/Twitter free-tier monthly write-cap headroom. Pass post_id to check one existing post; omit it to scan the forward queue (draft + scheduled, soonest first) and return only posts that have at least one issue.

    Annotations the server declares

    • title: Preflight Check
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.

Reads private data0
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write21
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out5
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 18:50 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
15
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
not observed
Authorization servers
not observed
Legacy HTTP+SSE transport
no
Dynamic client registration
not observed
Client ID metadata documents
not observed
  • HighThe whole server

    The endpoint served tools/list without credentials, and the list has 15 tools that change things (2 destructive, 1 irreversible, 12 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://viral-cat.com/api/mcp/mcp: tools/list returned 21 tools to an anonymous client, including cancel_post (destructive), update_brand (destructive), post_performance (irreversible), content_queue (write), campaign_summary (write) and 10 more

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

1 warn1 info
  • The server marks "post_performance" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "post_performance" says "post": a word for a change.

    What this check looks for
  • The description mentions "credentials".

    Matched text

    description: "...ublish or touch content, only refreshes credentials; safe to retry. Never exposes tokens."

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. Tools21 changed.~21 changedView diff

Install

Pinned to version 1.0.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http viral-cat https://viral-cat.com/api/mcp/mcp
claude mcp add --transport http viral-cat https://viral-cat.com/api/mcp/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__viral-cat__list_brands",
      "mcp__viral-cat__get_brand_details",
      "mcp__viral-cat__list_posts",
      "mcp__viral-cat__get_pipeline_summary",
      "mcp__viral-cat__list_connections",
      "mcp__viral-cat__preflight_check"
    ],
    "ask": [
      "mcp__viral-cat__post_performance",
      "mcp__viral-cat__content_queue",
      "mcp__viral-cat__campaign_summary",
      "mcp__viral-cat__diagnose_failures",
      "mcp__viral-cat__create_draft",
      "mcp__viral-cat__schedule_post",
      "mcp__viral-cat__retry_failed_posts",
      "mcp__viral-cat__reschedule_post",
      "mcp__viral-cat__refresh_post_metrics",
      "mcp__viral-cat__reconnect_platform",
      "mcp__viral-cat__upload_media",
      "mcp__viral-cat__create_media_upload",
      "mcp__viral-cat__create_brand"
    ],
    "deny": [
      "mcp__viral-cat__cancel_post",
      "mcp__viral-cat__update_brand"
    ]
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as viral-cat. Review it before you use it.

Capability tags behind 21 rules
  • mcp__viral-cat__list_brands untrusted-input
  • mcp__viral-cat__get_brand_details untrusted-input
  • mcp__viral-cat__list_posts untrusted-input
  • mcp__viral-cat__get_pipeline_summary untrusted-input
  • mcp__viral-cat__list_connections untrusted-input
  • mcp__viral-cat__post_performance untrusted-input, outbound
  • mcp__viral-cat__content_queue untrusted-input
  • mcp__viral-cat__campaign_summary untrusted-input
  • mcp__viral-cat__diagnose_failures untrusted-input
  • mcp__viral-cat__preflight_check untrusted-input
  • mcp__viral-cat__create_draft untrusted-input
  • mcp__viral-cat__schedule_post untrusted-input, outbound
  • mcp__viral-cat__cancel_post untrusted-input
  • mcp__viral-cat__retry_failed_posts untrusted-input, outbound
  • mcp__viral-cat__reschedule_post untrusted-input, outbound
  • mcp__viral-cat__refresh_post_metrics untrusted-input, outbound
  • mcp__viral-cat__reconnect_platform untrusted-input
  • mcp__viral-cat__upload_media untrusted-input
  • mcp__viral-cat__create_media_upload untrusted-input
  • mcp__viral-cat__create_brand untrusted-input
  • mcp__viral-cat__update_brand untrusted-input

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/com.viral-cat/viral-cat/verify?hash=418b02cafb42fd6885f90c5261a9ea969162d7e30e21f4b1dcac254ee2328f69" | jq -e '.changed == false'

Pinned hash 418b02cafb42, from Nov 14, 2026, 18:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/com.viral-cat/viral-cat/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for com.viral-cat/viral-cat
[![Anlyon Radar](https://radar.anlyon.com/badge/com.viral-cat/viral-cat.svg)](https://radar.anlyon.com/servers/com.viral-cat/viral-cat)

Source: MCP Registry. Synced Oct 21, 2026, 18:10 UTC.

ViralCat MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the ViralCat MCP server safe?

No scan can promise that. What its definitions show: 21 tools, 15 of which can change, send or delete something, and 2 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the ViralCat MCP server delete data?

Yes. 2 tools can delete, cancel or revoke something: cancel_post and update_brand.

Can the ViralCat MCP server do things that cannot be undone?

Yes. 1 tool can send, pay, publish or merge: post_performance.

What tools does the ViralCat MCP server have?

21 tools: 2 destructive, 1 irreversible, 12 write, 6 read. cancel_post, update_brand, post_performance, content_queue, campaign_summary, diagnose_failures, create_draft, schedule_post and 13 more.

Does the ViralCat MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the ViralCat MCP server send data to outside services?

5 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): post_performance, schedule_post, retry_failed_posts, reschedule_post and refresh_post_metrics.

Has the ViralCat MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Nov 1, 2026. Every change is on this page and in its Atom feed.