Ahel MCP server
ai.ahel/ahelahel.aistreamable-httpv1.0.1Scanned
Data and tools for your AI agent: company, people, SEO, social and EU register data. Pay per call.
11 tools: 9 change things and 2 only read.
- Reads private data · 1
- Reads text others write · 4
- Can send data out · 0
- 2 findings
Tools listed without sign-in. Remote posture
With Anlyon
9 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.
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 Ahel MCP server can do
The Ahel MCP server is published by ahel.ai, a domain the MCP Registry verified. It has 11 tools. It can create or change data (9 tools: run_capability, run_batch, run_recipe, capability_run_status and 5 more), but cannot delete data. 2 tools only read.
1 tool reads private data (capability_search); 4 tools bring in text other people can write (capability_search, run_capability, run_batch and 1 more).
The fixed checks matched 2 findings (annotation-conflict). 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 15, 2026, version 1.0.1.
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.
11 tools
Changes things 9Claude Code: ask
balanceThe workspace's prepaid balance, what running jobs have set aside, spending today and this month (for the Owner), any daily cap on this connection's key, and the top-up link.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 The workspace's prepaid balance, what running jobs have set aside, spending today and this month (for the Owner), any daily cap on this connection's key, and the top-up link. Use it when a run is refused for funds or before a large job. Free. NOT for: the price of one run, which is capability_inspect.
Annotations the server declares
- title: Read the prepaid balance
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
capability_run_statusRead a run that was still going: give execution_id from run_capability and results true for its rows.1 finding
Why it is in this group
- The name "capability_run_status" says "run": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "capability_run_status" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "capability_run_status" says "run": a word for a change.
Description, as the server wrote it Read a run that was still going: give execution_id from run_capability and results true for its rows. Answers the result in plain words, the charge and the rows; an answer with rows has results_page and csv_url: give the person those links instead of writing the rows out again. Long results answer a next_offset. include_raw true adds the full record and the provider's own answer, for debugging. Free. NOT for: finding an action, which is capability_search.
Annotations the server declares
- title: Read a run
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
catalog_requestAsk Ahel to add an action the catalog does not have, when capability_search finds nothing that fits.No findings
Why it is in this group
- The name "catalog_request" says "request": a word for a change.
Description, as the server wrote it Ask Ahel to add an action the catalog does not have, when capability_search finds nothing that fits. Give the job in plain words, one example input, and the most the person would pay for one run. Asking again adds to the same request. Free; runs and charges nothing. NOT for: finding what exists, which is capability_search.
Annotations the server declares
- title: Ask for a missing action
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
feedbackTell Ahel what went wrong or right: rate one run or request good or bad by its id, or with no id report anything confusing (results, charges, instructions, errors, a workaround you…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.
Description, as the server wrote it Tell Ahel what went wrong or right: rate one run or request good or bad by its id, or with no id report anything confusing (results, charges, instructions, errors, a workaround you needed). Report small annoyances too. Leave out private information and keys. Free. NOT for: reading a run, which is capability_run_status, or asking for a missing action, which is catalog_request.
Annotations the server declares
- title: Rate a run or report a problem
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
get_startedStart here when the person is new, asks what Ahel can do, or has no clear job yet.1 finding
Why it is in this group
- The name "get_started" says "get": a word for reading.
- The description's first verb is "start": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "get_started" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "get_started" says "get": a word for reading.
Description, as the server wrote it Start here when the person is new, asks what Ahel can do, or has no clear job yet. Answers what Ahel does in plain words, the ready-made recipes to offer first, how paying works, and a question to ask the person. Free; changes nothing. NOT for: a clear one-off job, which is capability_search.
Annotations the server declares
- title: Get started with Ahel
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
reviewRate how useful one finished run was, after using its result: useful, partly, not_useful or not_sure, with an optional reason (missing rows, wrong data, stale).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.
Description, as the server wrote it Rate how useful one finished run was, after using its result: useful, partly, not_useful or not_sure, with an optional reason (missing rows, wrong data, stale). Reviewing again replaces the earlier one. Free. NOT for: reporting a problem, which is feedback.
Annotations the server declares
- title: Review a run's usefulness
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
run_batchRun one action over many inputs in one call (1 to 100 items, each the action's inputs), such as a company lookup for 20 domains.Reads text others writeNo findings
Why it is in this group
- The name "run_batch" says "run": 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 Run one action over many inputs in one call (1 to 100 items, each the action's inputs), such as a company lookup for 20 domains. Priced the same per item as a single run, with no fee for batching, and only delivered results are charged; max_cost_usd caps the whole batch. Price one item with capability_inspect first and tell the person the total before running. The answer comes within about 45 seconds with each item's state, rows and charge; an item with rows has results_page and csv_url: give the person those links instead of writing the rows out again; items not finished by then finish when the same call is sent again (nothing runs or is charged twice). NOT for: one run, which is run_capability, or a ready-made job, which is run_recipe.
Annotations the server declares
- title: Run one action over many inputs
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
run_capabilityRun one action, after the person agreed to the price capability_inspect gave.Reads text others writeNo findings
Why it is in this group
- The name "run_capability" says "run": 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 Run one action, after the person agreed to the price capability_inspect gave. Give id, arguments and max_cost_usd from that price: a run that would cost more is refused with nothing charged. Only a delivered result is charged; a run that finds nothing costs nothing. It waits up to wait_seconds (default 20, at most 50; 0 answers at once) and answers the result in plain words with its rows; a longer run answers its execution_id for capability_run_status. An answer with rows has results_page and csv_url: give the person those links instead of writing the rows out again. idempotency_key makes a retry safe (the same key answers the same run, no new charge). Give tag (such as a customer name) to see usage and cost by tag in Activity. A list for people.email.find or companies.register.search or companies.emails.from_website: rows, or rows_text (paste up to 100 rows as text), instead of arguments; without confirm it answers the price of the list, with confirm true it runs it. NOT for: the same action over many inputs, which is run_batch, or a ready-made job, which is run_recipe.
Annotations the server declares
- title: Run an action
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
run_recipeRun a recipe: a ready-made job of several actions, such as local businesses of one kind in one place with their emails (or only those with weak reviews), new companies with contact…Reads text others writeNo findings
Why it is in this group
- The name "run_recipe" says "run": 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 Run a recipe: a ready-made job of several actions, such as local businesses of one kind in one place with their emails (or only those with weak reviews), new companies with contacts or companies with a reason to call now in any country (contracts, new registrations, funding or hiring, with a signal date, the person to reach and checked emails; Estonia adds year-on-year tax figures), a visibility report for a website (country, language, questions and optional competitors), a company list with industry, size and register entry, checked work emails for a list of people, a brief on a company and a person before a sales call, a pitch or an interview, with optional news and public posts, or companies that just raised money, use a named tool or won a public contract, each with the person to reach and a checked work email, or the social media accounts of one kind of business in one place, with when each started and how many followers it has. Visibility accepts English country/language names or codes. Reason-to-call groups Estonian contract lines by company and defaults to a 1,000 euro company total floor (minimum: 0 disables it). For emails with TikTok, include Instagram; TikTok returns no website. The first choice for a person who is not a developer. Without id: the list of recipes and their inputs. With id and inputs: the price of the whole job and a consent sentence; nothing runs. With max_cost_usd as well (once the person agrees): it runs under that total and answers the rows; a longer run answers a run_id, and run_id reads on. Only delivered results are charged. NOT for: one action alone, which is run_capability.
Annotations the server declares
- title: Run a ready-made recipe
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
Only reads 2Show 2 read-only tools
capability_inspectOne action by its id: its inputs, what it gives and its price rule.No findings
Why it is in this group
- The name "capability_inspect" says "inspect": a word for reading.
Description, as the server wrote it One action by its id: its inputs, what it gives and its price rule. With arguments (the person's real inputs) it also prices that exact run: the most it can cost, what it usually costs, the balance, and a consent sentence to read to the person. Nothing runs and nothing is charged. With rows or rows_text (up to 100, for people.email.find or companies.register.search or companies.emails.from_website) it prices a whole list. Use it before every paid run. For builders: endpoint_ids or exclude_endpoint_ids add the provider list those take. Free. NOT for: finding the id, which is capability_search, or running, which is run_capability.
Annotations the server declares
- title: Read and price an action
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
capability_searchFind the action for a job, by the person's own words ("work email of a person", "dentists in a city") or by an action id.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "capability_search" says "search": a word for reading.
What it can reach
- private-dataThe description's first sentence says "find" and "email": it reads mail.
- untrusted-inputThe description's first sentence says "find" and "email": it reads inbound email, which other people can write.
Description, as the server wrote it Find the action for a job, by the person's own words ("work email of a person", "dentists in a city") or by an action id. Use it first for any one-off data job. When a ready-made recipe fits the ask, the answer offers it first. Each result has its id, what it gives, whether it is ready and its price rule. Nothing fits: the answer says so and catalog_request asks Ahel to add it. Free; changes nothing. NOT for: the exact price of one run, which is capability_inspect.
Annotations the server declares
- title: Find an action
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- 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 outbound.
- Reads private data1
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write4
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out0
- 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.
Looked at Nov 15, 2026, 13:21 UTC, unchanged since Nov 15, 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
- 9
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://mcp.ahel.ai/.well-known/oauth-protected-resource
- Authorization servers
- https://glorious-elephant-19.authkit.app
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- yes
- Client ID metadata documents
- yes
- WarnThe whole server
The endpoint served tools/list without credentials, and the list has 9 tools that change things (9 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://mcp.ahel.ai: tools/list returned 11 tools to an anonymous client, including run_capability (write), run_batch (write), run_recipe (write), capability_run_status (write), catalog_request (write) and 4 more
Posture changes
- Nov 15, 2026
- Tools that change things, listed without credentials: 8 to 9
- Nov 8, 2026
- HTTP status of an anonymous initialize: 401 to 200
- tools/list served without credentials: no to yes
- Tools that change things, listed without credentials: 0 to 8
- Protocol version: none to 2025-06-18
- WWW-Authenticate on 401: yes to not observed
- resource_metadata in WWW-Authenticate: https://mcp.ahel.ai/.well-known/oauth-protected-resource to none
Findings
Matches of fixed checks, with the text they matched.
The server marks "capability_run_status" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "capability_run_status" says "run": a word for a change.
- Warnget_started
The server marks "get_started" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The name "get_started" says "get": a word for reading.
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools2 added, 3 removed, 9 changed.+2 added−3 removed~9 changedView diff
- Added:run_batchChanges things
- Added:run_recipeChanges things
- Removed:fetch
- Removed:my_tools
- Removed:search
- Changed:balanceChanges thingsdescription
- Changed:capability_inspectOnly readstitledescriptioninputSchemaannotations
- Changed:capability_run_statusChanges thingstitledescriptioninputSchemaannotations
- Changed:capability_searchOnly readstitledescriptioninputSchemaannotations
- Changed:catalog_requestChanges thingstitledescriptionannotations
- Changed:feedbackChanges thingstitledescriptioninputSchemaannotations
- Changed:get_startedChanges thingsdescription
- Changed:reviewChanges thingsdescription
- Changed:run_capabilityChanges thingstitledescriptioninputSchemaannotations
PostureRemote posture changed. Tools that change things, listed without credentials: 8 -> 9.View diff
- Tools that change things, listed without credentials: 8 to 9
PostureRemote posture changed. HTTP status of an anonymous initialize: 401 -> 200; tools/list served without credentials: no -> yes; Tools that change things, listed without credentials: 0 -> 8 (and 3 more).View diff
- HTTP status of an anonymous initialize: 401 to 200
- tools/list served without credentials: no to yes
- Tools that change things, listed without credentials: 0 to 8
- Protocol version: none to 2025-06-18
- WWW-Authenticate on 401: yes to not observed
- resource_metadata in WWW-Authenticate: https://mcp.ahel.ai/.well-known/oauth-protected-resource to none
Install
Pinned to version 1.0.1. Values in angle brackets are yours to fill in.
claude mcp add --transport http ahel https://mcp.ahel.aiclaude mcp add --transport http ahel https://mcp.ahel.aiRun 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.
{
"permissions": {
"allow": [
"mcp__ahel__capability_search",
"mcp__ahel__capability_inspect"
],
"ask": [
"mcp__ahel__run_capability",
"mcp__ahel__run_batch",
"mcp__ahel__run_recipe",
"mcp__ahel__capability_run_status",
"mcp__ahel__catalog_request",
"mcp__ahel__feedback",
"mcp__ahel__balance",
"mcp__ahel__review",
"mcp__ahel__get_started"
],
"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 ahel. Review it before you use it.
Capability tags behind 4 rules
- mcp__ahel__capability_search private-data, untrusted-input
- mcp__ahel__run_capability untrusted-input
- mcp__ahel__run_batch untrusted-input
- mcp__ahel__run_recipe 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/ai.ahel/ahel/verify?hash=145f1a610eee87134d52cd36a6289bef345cb56d94f8f0e32da9af0291eac32d" | jq -e '.changed == false'Pinned hash 145f1a610eee, from Nov 15, 2026, 19:17 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/ai.ahel/ahel/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/ai.ahel/ahel)Source: MCP Registry · repository. Synced Nov 2, 2026, 13:17 UTC.
Ahel MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Ahel MCP server safe?
- No scan can promise that. What its definitions show: 11 tools, 9 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 Ahel MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the Ahel 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 Ahel MCP server have?
- 11 tools: 9 write, 2 read. run_capability, run_batch, run_recipe, capability_run_status, catalog_request, feedback, balance, review and 3 more.
Does the Ahel MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Ahel MCP server send data to outside services?
- None of its tools sends, posts or publishes, or takes a recipient, channel or URL.
Has the Ahel MCP server changed its tools?
- Yes, once since Anlyon Radar started tracking it, most recently on Nov 15, 2026. Every change is on this page and in its Atom feed.