Skip to content

Toolbox MCP server

com.allthepossibles/toolboxallthepossibles.comstreamable-httpv1.2.0Scanned

Small tools for agent builders: AI-writing tells, tools/list token cost, CLAUDE.md lint, and more.

14 tools: 10 change things and 4 only read.

Tools: 0 delete, 0 irreversible, 10 write, 4 read.
  • Reads private data · 0
  • Reads text others write · 8
  • Can send data out · 8
  • 3 findings, 2 high

Tools listed without sign-in. Remote posture

With Anlyon

10 tools here change things. 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 Toolbox MCP server can do

The Toolbox MCP server is published by allthepossibles.com, a domain the MCP Registry verified. It has 14 tools. It can create or change data (10 tools: check_tells, manifest_cost, lint_agent_md, mcp_install_snippets and 6 more), but cannot delete data. 4 tools only read.

8 tools bring in text other people can write (manifest_cost, mcp_install_snippets, check_ai_crawlers and 5 more); 8 tools can send data out of your system (manifest_cost, mcp_install_snippets, check_ai_crawlers and 5 more).

The fixed checks matched 3 findings, 2 of them high severity (annotation-conflict and sensitive-path). 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 11, 2026, version 1.2.0.

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.

14 tools

Changes things 10Claude Code: ask
  • check_sitemapCheck a site's sitemaps: discovers them from robots.txt, /sitemap.xml and /sitemap_index.xml, reads index files one level deep (20 children max), counts URLs per file and per host,…Reads text others writeCan send data outNo 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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Check a site's sitemaps: discovers them from robots.txt, /sitemap.xml and /sitemap_index.xml, reads index files one level deep (20 children max), counts URLs per file and per host, summarizes lastmod, and fetches a sample of URLs (HEAD, then GET) to count the ones that are not 200. Flags a missing sitemap, no robots.txt entry, over 50,000 URLs, other hosts, http URLs on https sites and missing lastmod.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • check_tellsCheck text for AI-writing tells (em-dashes, 'not just X but Y', rule-of-three lists, stock vocabulary like delve or seamlessly, hedge words, smart quotes, hidden Unicode) and retur…No 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.
    Description, as the server wrote it
    Check text for AI-writing tells (em-dashes, 'not just X but Y', rule-of-three lists, stock vocabulary like delve or seamlessly, hedge words, smart quotes, hidden Unicode) and return a score, each tell with its position, and a cleaned copy. Run it on anything you are about to publish on a human's behalf.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
  • decode_jwtDecode a JWT's header and payload without verifying the signature (it never takes a secret).No 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.
    Description, as the server wrote it
    Decode a JWT's header and payload without verifying the signature (it never takes a secret). Returns the claims with iat, nbf and exp as ISO times, a plain-text expiresIn, issuer, subject and audience, and findings: alg none, HS* with a URL-like kid, missing or distant exp, expired, nbf or iat in the future, payload over 4 KB, sensitive field names, and jku or x5u pointing off-site.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
  • json_to_schemaTurn a JSON sample (or an array of samples of the same shape) into a JSON Schema draft 2020-12 and a TypeScript type.No 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.
    Description, as the server wrote it
    Turn a JSON sample (or an array of samples of the same shape) into a JSON Schema draft 2020-12 and a TypeScript type. Infers required vs optional keys across samples, nullable fields, integer vs number, date/email/uri/uuid formats, and small enums. Use it to draft an MCP tool inputSchema from an example payload.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
  • lint_agent_mdLint a CLAUDE.md, AGENTS.md or MEMORY.md: line count against Claude Code's 200-line memory index cap, estimated tokens, duplicate lines, contradicting rules, leftover TODO markers,…No 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.
    Description, as the server wrote it
    Lint a CLAUDE.md, AGENTS.md or MEMORY.md: line count against Claude Code's 200-line memory index cap, estimated tokens, duplicate lines, contradicting rules, leftover TODO markers, hidden characters, empty headings.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
  • lint_mcp_toolsLint the tools an MCP server lists, from the model's side of the wire: names that break Claude's ^[a-zA-Z0-9_-]{1,64}$ rule, duplicate names, missing or near-identical descriptions…Reads text others writeCan send data outNo 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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Lint the tools an MCP server lists, from the model's side of the wire: names that break Claude's ^[a-zA-Z0-9_-]{1,64}$ rule, duplicate names, missing or near-identical descriptions, instruction-like text in a description, arguments with no description, required arguments missing from properties, destructive-sounding tools with no destructiveHint, and unbounded free-text arguments on them. Pass a live Streamable HTTP endpoint or a tools array.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • manifest_costEstimate what an MCP server's tools/list costs in tokens on every session start, per tool, with the three most expensive tools and flags for oversized descriptions, base64 blobs an…Reads text others writeCan send data outNo 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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Estimate what an MCP server's tools/list costs in tokens on every session start, per tool, with the three most expensive tools and flags for oversized descriptions, base64 blobs and echoed schemas. Pass a live Streamable HTTP endpoint or a tools array.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • mcp_install_snippetsGenerate install snippets for an MCP endpoint (Claude Code, Claude Desktop, Cursor, Codex, VS Code, curl) and probe the endpoint once to confirm it answers initialize and tools/lis…Reads text others writeCan send data out1 finding

    Why it is in this group

    • The name "mcp_install_snippets" says "install": a word for a change.
    • 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 input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Warn

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

      Matched text

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

      What this check looks for
    Description, as the server wrote it
    Generate install snippets for an MCP endpoint (Claude Code, Claude Desktop, Cursor, Codex, VS Code, curl) and probe the endpoint once to confirm it answers initialize and tools/list.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • scrub_envScrub a .env file into a safe .env.example: every value is blanked or replaced with a placeholder that keeps its shape (postgres://USER:PASSWORD@HOST/DB).2 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.

    Findings

    • High

      The description refers to ".env" and ".env.example".

      Matched text

      description: "Scrub a .env file into a safe .env.example: every va..."

      What this check looks for
    • High

      The input schema at properties.text.description refers to ".env".

      Matched text

      inputSchema.properties.text.description: "The .env file contents. Free tier: 20,000 chars."

      What this check looks for
    Description, as the server wrote it
    Scrub a .env file into a safe .env.example: every value is blanked or replaced with a placeholder that keeps its shape (postgres://USER:PASSWORD@HOST/DB). Returns the example, the secrets found (AWS, GitHub, Stripe, Slack, Google, OpenAI, Anthropic, Resend, SendGrid, Twilio, Cloudflare, JWTs, private keys, database URLs, high-entropy strings) by key, kind and line with only the first 4 characters and the length, and findings for duplicate keys, non-UPPER_SNAKE names, unbalanced quotes, export prefixes and trailing spaces.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
  • trace_redirectsTrace a URL's redirect chain hop by hop (up to 10 hops, 8 s each) with status, location, timing and scheme per hop.Reads text others writeCan send data outNo 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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Trace a URL's redirect chain hop by hop (up to 10 hops, 8 s each) with status, location, timing and scheme per hop. Flags chains over two redirects, loops, HTTPS-to-HTTP downgrades, 302s used for www or https canonicalisation, dropped paths or query strings, cross-domain moves, meta-refresh or JS redirects on a 200, and a final page that is not 200.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
Only reads 4Show 4 read-only tools
  • check_ai_crawlersRead a site's robots.txt and report which AI crawlers (GPTBot, ClaudeBot, Google-Extended, PerplexityBot, CCBot, Bytespider and more) are allowed, blocked or partially blocked; che…Reads text others writeCan send data outNo 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 input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Claude Code: ask, because it is tagged outbound.

    Description, as the server wrote it
    Read a site's robots.txt and report which AI crawlers (GPTBot, ClaudeBot, Google-Extended, PerplexityBot, CCBot, Bytespider and more) are allowed, blocked or partially blocked; check for llms.txt and draft one from the site's metadata and sitemap.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • check_headersFetch a URL once and grade its response headers: HSTS, Content-Security-Policy, X-Content-Type-Options, frame control, Referrer-Policy, Permissions-Policy, Cache-Control, cookie Se…Reads text others writeCan send data outNo findings

    Why it is in this group

    • The description's first verb is "fetch": 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 "fetch" and "url": it reads web pages, which anyone can write.
    • untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Claude Code: ask, because it is tagged outbound.

    Description, as the server wrote it
    Fetch a URL once and grade its response headers: HSTS, Content-Security-Policy, X-Content-Type-Options, frame control, Referrer-Policy, Permissions-Policy, Cache-Control, cookie Secure/HttpOnly/SameSite flags, and server version leaks. Returns a grade, a score and the exact header to add for each gap.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • check_page_metaFetch a URL and report what crawlers and link unfurlers see: title and meta description with lengths, canonical, lang, viewport, robots meta, h1s, Open Graph and Twitter card tags,…Reads text others writeCan send data outNo findings

    Why it is in this group

    • The description's first verb is "fetch": 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 "fetch" and "url": it reads web pages, which anyone can write.
    • untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Claude Code: ask, because it is tagged outbound.

    Description, as the server wrote it
    Fetch a URL and report what crawlers and link unfurlers see: title and meta description with lengths, canonical, lang, viewport, robots meta, h1s, Open Graph and Twitter card tags, favicon, JSON-LD types, the resulting preview card, and findings for anything missing or too long.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: true
  • explain_cronExplain a five-field cron expression in plain English, list the next runs in UTC, and warn about the cases Cloudflare Workers, GitHub Actions and crontab handle differently (nickna…No findings

    Why it is in this group

    • The name "explain_cron" says "explain": a word for reading.
    Description, as the server wrote it
    Explain a five-field cron expression in plain English, list the next runs in UTC, and warn about the cases Cloudflare Workers, GitHub Actions and crontab handle differently (nicknames, six fields, day-of-month plus day-of-week, sub-5-minute intervals, day 31).

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false

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 write8
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out8
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, 19:24 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
10
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
  • WarnThe whole server

    The endpoint served tools/list without credentials, and the list has 10 tools that change things (10 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://toolbox.allthepossibles.com/mcp: tools/list returned 14 tools to an anonymous client, including check_tells (write), manifest_cost (write), lint_agent_md (write), mcp_install_snippets (write), json_to_schema (write) and 5 more

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

2 high1 warn
  • The description refers to ".env" and ".env.example".

    Matched text

    description: "Scrub a .env file into a safe .env.example: every va..."

    What this check looks for
  • The input schema at properties.text.description refers to ".env".

    Matched text

    inputSchema.properties.text.description: "The .env file contents. Free tier: 20,000 chars."

    What this check looks for
  • The server marks "mcp_install_snippets" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

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

    What this check looks for

History

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

  1. Tools1 changed.~1 changedView diff

Install

Pinned to version 1.2.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http toolbox https://toolbox.allthepossibles.com/mcp
claude mcp add --transport http toolbox https://toolbox.allthepossibles.com/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__toolbox__explain_cron"
    ],
    "ask": [
      "mcp__toolbox__check_tells",
      "mcp__toolbox__manifest_cost",
      "mcp__toolbox__lint_agent_md",
      "mcp__toolbox__mcp_install_snippets",
      "mcp__toolbox__check_ai_crawlers",
      "mcp__toolbox__check_headers",
      "mcp__toolbox__check_page_meta",
      "mcp__toolbox__json_to_schema",
      "mcp__toolbox__scrub_env",
      "mcp__toolbox__trace_redirects",
      "mcp__toolbox__check_sitemap",
      "mcp__toolbox__decode_jwt",
      "mcp__toolbox__lint_mcp_tools"
    ],
    "deny": []
  }
}

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 toolbox. Review it before you use it.

Capability tags behind 8 rules
  • mcp__toolbox__manifest_cost untrusted-input, outbound
  • mcp__toolbox__mcp_install_snippets untrusted-input, outbound
  • mcp__toolbox__check_ai_crawlers untrusted-input, outbound
  • mcp__toolbox__check_headers untrusted-input, outbound
  • mcp__toolbox__check_page_meta untrusted-input, outbound
  • mcp__toolbox__trace_redirects untrusted-input, outbound
  • mcp__toolbox__check_sitemap untrusted-input, outbound
  • mcp__toolbox__lint_mcp_tools untrusted-input, outbound

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.allthepossibles/toolbox/verify?hash=69f475531c252cd0bd0b30b0d9fd2029eef9d87076e3a7dae927bebbd2ede145" | jq -e '.changed == false'

Pinned hash 69f475531c25, from Nov 11, 2026, 14:17 UTC.

Watch

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

https://radar.anlyon.com/servers/com.allthepossibles/toolbox/changes.xml

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

Badge

Anlyon Radar badge for com.allthepossibles/toolbox
[![Anlyon Radar](https://radar.anlyon.com/badge/com.allthepossibles/toolbox.svg)](https://radar.anlyon.com/servers/com.allthepossibles/toolbox)

Source: MCP Registry. Synced Oct 30, 2026, 00:17 UTC.

Toolbox MCP server: questions and answers

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

Is the Toolbox MCP server safe?

No scan can promise that. What its definitions show: 14 tools, 10 of which can change, send or delete something, and 3 findings (2 high) from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the Toolbox MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

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

No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.

What tools does the Toolbox MCP server have?

14 tools: 10 write, 4 read. check_tells, manifest_cost, lint_agent_md, mcp_install_snippets, json_to_schema, scrub_env, trace_redirects, check_sitemap and 6 more.

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

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

Does the Toolbox MCP server send data to outside services?

8 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): manifest_cost, mcp_install_snippets, check_ai_crawlers, check_headers, check_page_meta, trace_redirects and 2 more.

Has the Toolbox MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Oct 30, 2026. Every change is on this page and in its Atom feed.