Lawstronaut MCP server
com.lawstronaut/lawstronaut-mcplawstronaut.comstreamable-httpv2.0.0Scanned
Connect AI to millions of laws and court cases with the Lawstronaut MCP.
18 tools: 2 change things and 16 only read.
- Reads private data · 8
- Reads text others write · 18
- Can send data out · 4
- 23 findings
Tools listed without sign-in. Remote posture
With Anlyon
2 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 Lawstronaut MCP server can do
The Lawstronaut MCP server is published by lawstronaut.com, a domain the MCP Registry verified. It has 18 tools. It can create or change data (2 tools: horizon_scan and health_check), but cannot delete data. 16 tools only read.
8 tools read private data (list_jurisdictions, list_documents, list_full_text and 5 more); 18 tools bring in text other people can write (list_jurisdictions, list_domains, list_subdomains and 15 more); 4 tools can send data out of your system (list_documents, list_full_text, list_markdown and 1 more). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.
The fixed checks matched 23 findings (credentials-mention and long-description). 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 9, 2026, version 2.0.0.
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.
18 tools
Changes things 2Claude Code: ask
health_checkVerify that the Lawstronaut MCP server and its backing services (database, cache) are reachable and responding.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 Verify that the Lawstronaut MCP server and its backing services (database, cache) are reachable and responding. Returns the overall status, per-service status, server version, and a UTC timestamp. When to use: Call at the start of a session to confirm connectivity before running any data tools, or when a previous tool call returned an unexpected error and you want to diagnose whether the service itself is down. When not: Do not call on every tool invocation — use it only for diagnostic purposes. It does not return legal data. Parameters: Takes none — invoke with {}. No filters, pagination, or auth-scoped data. Behavior: Read-only: one lightweight SELECT 1 against the database and one PING against the cache. Does not create, update, or delete any records. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. If the database is unreachable the overall status will be 'degraded'. If only the cache is unreachable the overall status will be 'degraded'. Both must be healthy for status 'ok'. Returns: status ('ok' | 'degraded'), version string, utc_timestamp, and a services object with db and cache each reporting their own status ('ok' | 'error') plus optional latency_ms.
Annotations the server declares
- title: Health check
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
horizon_scanScans recent legal documents across jurisdictions.Reads text others write1 finding
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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Scans recent legal documents across jurisdictions. When to use: When you need to find legal content for a specific jurisdiction and optional filters. When not: Do not use this to read a document body or download a file. For plain text use list_full_text. For structured Markdown use list_markdown. For a complete citation bundle use evidence_pull. For a temporary source-file URL use document_presigned_url. Do not call without an iso from list_jurisdictions. Skip if you already have this scan for the same iso and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions (e.g. BE, or US_CA for a state). issuing_authority, type_of_authority. Topic, since, date_field, status, include_repealed, max_per_jurisdiction are optional. Empty data means the iso is outside the subscription, or no document matched. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and document data. Documents with total_links of 0 are omitted. Returns: A list of documents matching the filters, with pagination information.
Annotations the server declares
- title: Horizon scan
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
Only reads 16Show 16 read-only tools
document_presigned_urlReturns a temporary signed URL to the original source file (PDF or other stored file) for one document.Reads private dataReads text others write1 finding
Why it is in this group
- The description's first verb is "return": a word for reading.
What it can reach
- private-dataThe name "document_presigned_url" says "document": it reads documents.
- private-dataThe description's first sentence says "return" and "file": it reads files.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe name "document_presigned_url" says "url": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "return" and "url": it reads web pages, which anyone can write.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
Description, as the server wrote it Returns a temporary signed URL to the original source file (PDF or other stored file) for one document. The URL expires. This tool does not return the document body or a citation bundle. When to use: When you need to open or download the original file. Examples: 'give me the PDF for uuid …', 'link to the source file so I can download it'. Use the uuid from list_documents, search, or horizon_scan. When not: For plain text of the body, use list_full_text (example: 'quote the operative wording'). For structured Markdown, use list_markdown (example: 'keep the article headings'). For a complete citation bundle — metadata, permanent legal link, citation template, body, and this temporary URL together — use evidence_pull (example: 'show me the source'). For a permanent citation link without a file download, use the legal_link from evidence_pull or list_documents. Do not call without a uuid. Skip if you already have an unexpired URL for that uuid. Parameters: uuid (required) is the document UUID. There is no iso parameter. The response includes expires_at; fetch the URL before it expires. A missing file or an archived document returns an error rather than a URL. Behavior: Read-only. Looks up the stored file and signs a temporary URL. Does not create, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached. Each call signs a new URL. expires_at is when that URL stops working. The URL is not a permanent legal link. Returns: One object: uuid, url (temporary source-file URL), and expires_at. url is empty when no file could be signed.
Annotations the server declares
- title: Get source url
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
document_versionRetrieves a specific document with a specific version When to use: The user wants to find a specific document with a specific version When not: The user wants to find a specific do…Reads private dataReads text others write1 finding
Why it is in this group
- The description's first verb is "retrieve": a word for reading.
What it can reach
- private-dataThe name "document_version" says "document": it reads documents.
- private-dataThe description's first sentence says "retrieve" and "document": it reads documents.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Retrieves a specific document with a specific version When to use: The user wants to find a specific document with a specific version When not: The user wants to find a specific document with a specific version but the document is not found Parameters: uuid (required) is the UUID of the document, version (required) is the version of the document Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and document data. Documents with total_links of 0 are omitted. Returns: A document with the specific version
Annotations the server declares
- title: Get document with version
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
evidence_pullReturns one complete citation bundle for a single document: metadata, a permanent legal_link, a citation_template, the body, and a temporary source-file URL when one exists.Reads private dataReads text others write2 findings
Why it is in this group
- The description's first verb is "return": a word for reading.
What it can reach
- private-dataThe description's first sentence says "return" and "document": it reads documents.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "return" and "url": it reads web pages, which anyone can write.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Info
The description is 2,303 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "..._template, and, when requested, content_format plus content, and source_file_url."
Description, as the server wrote it Returns one complete citation bundle for a single document: metadata, a permanent legal_link, a citation_template, the body, and a temporary source-file URL when one exists. When to use: When one call must support a citation or an audit. Examples: 'show me the source for this Irish statute', 'cite uuid … with its authority, date, link, and text'. Prefer this over calling the body and file tools separately. When not: For plain text alone, use list_full_text (example: 'quote the operative wording'). For structured Markdown alone, use list_markdown (example: 'keep the article headings'). For only a temporary URL to the original PDF or source file, use document_presigned_url (example: 'download the original file'). To browse or search without the bundle, use list_documents, search, or horizon_scan. For one historical version, use document_version. Do not call without an iso from list_jurisdictions and the document uuid. Skip if you already have this bundle for the same iso and uuid. Parameters: iso (required) is a jurisdiction code from list_jurisdictions and is uppercased (e.g. IE, NL, US_CA). uuid (required) is the document UUID from list_documents, search, or horizon_scan. include_text defaults to true and fetches the body. include_source_url defaults to true and adds a temporary signed URL to the original file when one exists. text_format defaults to 'markdown' (structured Markdown); 'plain' returns plain text instead. Set include_text or include_source_url to false only when that part of the bundle is unnecessary. Behavior: Read-only. Reads the document record, optionally the stored body, and optionally signs a temporary source-file URL. Does not create, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached. The source-file URL is temporary: source_file_note says to fetch it promptly, and legal_link is the permanent citation URL. source_file_url is null when no original file is stored. An archived document returns an archived error. Returns: One citation bundle: uuid, version, metadata, legal_link, citation_template, and, when requested, content_format plus content, and source_file_url.
Annotations the server declares
- title: Evidence pull
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_authority_typesLists the authority types available through this tool, with metadata such as jurisdiction and portal name.Reads text others write1 finding
Why it is in this group
- The name "list_authority_types" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists the authority types available through this tool, with metadata such as jurisdiction and portal name. When to use: When you need authority types for a known iso. Pass optional portal name or authority type only to narrow that list. When not: Do not call without an iso from list_jurisdictions. Skip if you already have this authority type list for the same iso and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions and is uppercased before the query (e.g. BE, or US_CA for a state). portal_name and authority_type are optional. authority_type must match one authority type exactly. With limit, next_id parameters to paginate the results. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and authority type data. Authority types with total_links of 0 are omitted. Returns: A list of authority types matching the filters, with pagination information.
Annotations the server declares
- title: List type of authority
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_categoriesLists categories under one subdomain, with the parent domain and subdomain names.Reads text others write1 finding
Why it is in this group
- The name "list_categories" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists categories under one subdomain, with the parent domain and subdomain names. This is the third taxonomy step. When to use: When you already have a subdomain_id from list_subdomains. Examples: 'categories under subdomain A.1', 'find a category named jurisdiction'. Pass category_name only to narrow by name. When not: Do not use this for subcategories; use list_subcategories with the returned category_id. Do not use it for law types; use list_law_types. For the parent lists use list_domains and list_subdomains. For documents use list_documents or search. Do not call without a subdomain_id from list_subdomains. Skip if you already have this category list for the same subdomain_id and category_name. Parameters: subdomain_id is the subdomain reference from list_subdomains, such as A.1. A call without it, or with an unknown value, returns Subdomain not found. category_name is an optional case-insensitive substring of the category name; the parameter is category_name, not category. limit defaults to 20 and is capped at 100. next_id is the numeric id from pagination.next_id, not category_id. Empty data means that subdomain has no matching category. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads the current category table. total_count is the full match count, not the page size. The input marks subdomain_id optional, but the query returns Subdomain not found when it is missing. Returns: A page of {id, domain, subdomain, category_id, category_name}. domain and subdomain are display names, not reference codes. Example category_id "A.1.1", category_name "General Private International Law". Pass category_id, not id, to list_subcategories.
Annotations the server declares
- title: List categories
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_documentsLists document metadata in a jurisdiction (title, dates, source, classification) with filters and pagination.Reads private dataReads text others writeCan send data out2 findings
Why it is in this group
- The name "list_documents" says "list": a word for reading.
What it can reach
- private-dataThe name "list_documents" says "list" and "document": it reads documents.
- private-dataThe description's first sentence says "list" and "document": it reads documents.
- private-dataThe input names a file (input.file_data_only), and the tool reads: it reads files.
- 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Info
The description is 2,395 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...Returns: A list of documents matching the filters, with pagination information."
Description, as the server wrote it Lists document metadata in a jurisdiction (title, dates, source, classification) with filters and pagination. It does not return the document body, a citation bundle, or a source-file URL. When to use: When you need to find or browse documents before reading them. Example: 'which Belgian employment statutes mention night work?' Use the returned uuid with a body or citation tool. When not: Do not use this for the body or the file. For plain text, use list_full_text (example: 'quote the operative wording'). For structured Markdown, use list_markdown (example: 'keep the article headings'). For a complete citation bundle — metadata, permanent legal link, citation template, body, and a temporary source-file URL — use evidence_pull (example: 'show me the source'). For only a temporary URL to the original file, use document_presigned_url (example: 'download the PDF'). Do not call without an iso from list_jurisdictions. Skip if you already have this document list for the same iso and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions (e.g. BE, or US_CA for a state). version, section_title, date_of_decision, file_data_only, issuing_authority, type_of_authority, source_identifier, source_secondary_identifier, tag are optional, section_title is a case-insensitive substring of the section title, date_of_decision is a date, file_data_only is a boolean, issuing_authority is a case-insensitive substring of the issuing authority, type_of_authority is a case-insensitive substring of the type of authority, source_identifier is a case-insensitive substring of the source identifier, source_secondary_identifier is a case-insensitive substring of the source secondary identifier, tag is a case-insensitive substring of the tag. With limit, next_id parameters to paginate the results. Empty data means the iso is outside the subscription, or no document matched. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and document data. Documents with total_links of 0 are omitted. Returns: A list of documents matching the filters, with pagination information.
Annotations the server declares
- title: List documents
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_domainsLists the top-level legal domains (id, domain_id, domain name).Reads text others write1 finding
Why it is in this group
- The name "list_domains" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists the top-level legal domains (id, domain_id, domain name). This is the first taxonomy step. It does not return subdomains, categories, documents, or jurisdictions. When to use: When you need the domain list or a name match. Examples: 'what legal domains exist?', 'find the domain for private law'. Pass domain only to narrow by name. When not: Do not use this for children of a domain. For subdomains use list_subdomains with the returned domain_id. For categories use list_categories. For subcategories use list_subcategories. For law types use list_law_types. For documents use list_documents or search. For jurisdictions use list_jurisdictions. Skip if you already have this domain list for the same domain filter. Parameters: domain is an optional case-insensitive substring of the domain name. There is no parent id and no iso. limit defaults to 20 and is capped at 100. next_id is the numeric id from the previous pagination.next_id; omit it on the first page. A null next_id means there is no further page. Empty data means no domain name matched. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads the current domain table. total_count is the full match count, not the page size. Results are not limited to a jurisdiction. Returns: A page of {id, domain_id, domain}. Example: {"id":1,"domain_id":"A","domain":"Private law"}. Pass domain_id, not id, to list_subdomains.
Annotations the server declares
- title: List domains
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_full_textReturns the plain-text body of documents (id, uuid, version, text).Reads private dataReads text others writeCan send data out2 findings
Why it is in this group
- The name "list_full_text" says "list": a word for reading.
What it can reach
- private-dataThe description's first sentence says "return" and "document": it reads documents.
- 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Info
The description is 2,220 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...n text. Empty data means nothing matched, or the iso is outside the subscription."
Description, as the server wrote it Returns the plain-text body of documents (id, uuid, version, text). This is the plain-text read: wording only, with no Markdown headings or lists, and no citation metadata. When to use: When you need the body as plain text. Examples: 'quote the operative wording of this Belgian statute', 'give me the raw text of uuid …'. Pass uuid when you already have one document; otherwise filter by title and paginate. When not: For structured Markdown that keeps headings and lists, use list_markdown. For a complete citation bundle in one call — metadata, permanent legal link, citation template, body, and a temporary source-file URL — use evidence_pull (example: 'show me the source'). For only a temporary URL to the original PDF or source file, use document_presigned_url (example: 'download the original file'). To browse titles and dates without the body, use list_documents or search. For one historical version's metadata plus both bodies, use document_version. Do not call without an iso from list_jurisdictions. Skip if you already have this plain text for the same iso, uuid, and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions and is uppercased before the query (e.g. BE, or US_CA for a state). uuid is optional; when set, the call targets that document. title is an optional case-insensitive substring. with_empty_data defaults to false and omits documents that have no stored plain text; true includes them with text null. limit defaults to 20 and is capped at 100. include_total adds pagination.total_count when true. Other optional filters match the input schema and are combined with AND. Behavior: Read-only SELECT plus a read of the stored text. Does not create, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription data and the stored plain text. An archived uuid returns an archived error instead of the body. Returns: A page of {id, uuid, version, text} where text is plain text. Empty data means nothing matched, or the iso is outside the subscription.
Annotations the server declares
- title: Get document text
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_issuing_authoritiesLists the issuing authorities available through this tool, with metadata such as jurisdiction and issuing authority.Reads text others write1 finding
Why it is in this group
- The name "list_issuing_authorities" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists the issuing authorities available through this tool, with metadata such as jurisdiction and issuing authority. When to use: When you need issuing authorities for a known iso. Pass optional portal name or issuing authority only to narrow that list. When not: Do not call without an iso from list_jurisdictions. Skip if you already have this issuing authority list for the same iso and filters. Empty data means the iso is outside the subscription, or no issuing authority matched. Parameters: iso (required) is a jurisdiction code from list_jurisdictions (e.g. BE, or US_CA for a state). portal_name and issuing_authority are optional. issuing_authority must match one issuing authority exactly. With limit, next_id parameters to paginate the results. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and issuing authority data. Issuing authorities with total_links of 0 are omitted. Returns: A list of issuing authorities matching the filters, with pagination information.
Annotations the server declares
- title: List issuing authorities
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_jurisdictionsList every country and state in the authenticated subscription's Lawstronaut legal corpus (name, iso, type, documents_count) — not the global unfiltered catalog, and not legal docu…Reads private dataReads text others write1 finding
Why it is in this group
- The name "list_jurisdictions" says "list": a word for reading.
What it can reach
- private-dataThe description's first sentence says "list" and "document": it reads documents.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The description mentions "credential".
What this check looks forMatched text
description: "...know which iso codes are valid for this credential. When not: Skip if you already have thi..."
Description, as the server wrote it List every country and state in the authenticated subscription's Lawstronaut legal corpus (name, iso, type, documents_count) — not the global unfiltered catalog, and not legal documents or full text. When to use: Call FIRST, before any jurisdiction-scoped legal query, so you know which iso codes are valid for this credential. When not: Skip if you already have this list from earlier in the session. Do not use this to fetch documents, markdown, source files, or portals. Parameters: Takes none — invoke with {}. Do not pass iso, country, limit, offset, or other filters; they are not accepted and would not change the result. There is no page size or default filter. Empty data means the subscription includes no jurisdictions. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and portal data. Returns: A list of jurisdictions matching the filters, with pagination information.
Annotations the server declares
- title: List jurisdictions
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_law_typesLists law types under one subcategory, with the parent category and subcategory names.Reads text others write1 finding
Why it is in this group
- The name "list_law_types" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists law types under one subcategory, with the parent category and subcategory names. This is the last taxonomy step. It does not return documents. When to use: When you already have a subcategory_id from list_subcategories. Examples: 'law types under subcategory A.1.1.1', 'find law types named statute'. Pass law_type only to narrow by name. When not: Do not use this to list subcategories, categories, domains, or documents. For those use list_subcategories, list_categories, list_domains, list_documents, or search. Do not call without a subcategory_id from list_subcategories. Skip if you already have this law type list for the same subcategory_id and law_type filter. Parameters: subcategory_id (required) is the subcategory reference from list_subcategories. It must be exactly four dot-separated segments, such as A.1.1.1, and is uppercased before the query. law_type is an optional case-insensitive substring of the law type name. A subcategory_id with the wrong shape is rejected before the query. limit defaults to 20 and is capped at 100. next_id is the numeric id from pagination.next_id, not subcategory_id. Empty data means that subcategory has no matching law type. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads the current law type table. total_count is the full match count, not the page size. There is no further taxonomy tool after this one. Returns: A page of {id, category_id, category_name, subcategory_id, subcategory_name, law_type_id, law_type}. id and law_type_id are the same numeric id. Example law_type "Statute".
Annotations the server declares
- title: List law types
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_markdownReturns the structured Markdown body of documents (id, uuid, version, markdown).Reads private dataReads text others writeCan send data out2 findings
Why it is in this group
- The name "list_markdown" says "list": a word for reading.
What it can reach
- private-dataThe description's first sentence says "return" and "document": it reads documents.
- 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Info
The description is 2,290 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...rkdown. Empty data means nothing matched, or the iso is outside the subscription."
Description, as the server wrote it Returns the structured Markdown body of documents (id, uuid, version, markdown). Headings and lists stay intact. This is not plain text, not a citation bundle, and not a file URL. When to use: When you need the body as structured Markdown. Examples: 'show this regulation with its article headings', 'give me Markdown I can render for uuid …'. Pass uuid when you already have one document; otherwise filter by title and paginate. When not: For plain text with no Markdown structure, use list_full_text (example: 'quote the operative wording'). For a complete citation bundle in one call — metadata, permanent legal link, citation template, body, and a temporary source-file URL — use evidence_pull (example: 'show me the source'). For only a temporary URL to the original PDF or source file, use document_presigned_url (example: 'download the original file'). To browse titles and dates without the body, use list_documents or search. For one historical version's metadata plus both bodies, use document_version. Do not call without an iso from list_jurisdictions. Skip if you already have this Markdown for the same iso, uuid, and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions and is uppercased before the query (e.g. BE, or US_CA for a state). uuid is optional; when set, the call targets that document. title is an optional case-insensitive substring. with_empty_data defaults to false and omits documents that have no stored Markdown; true includes them with markdown null. limit defaults to 20 and is capped at 100. include_total adds pagination.total_count when true. Other optional filters match the input schema and are combined with AND. Behavior: Read-only SELECT plus a read of the stored Markdown. Does not create, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription data and the stored Markdown. An archived uuid returns an archived error instead of the body. Returns: A page of {id, uuid, version, markdown} where markdown is structured Markdown. Empty data means nothing matched, or the iso is outside the subscription.
Annotations the server declares
- title: Get document markdown
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_portalsLists the source portals available through this tool, with metadata such as jurisdiction and language.Reads text others write1 finding
Why it is in this group
- The name "list_portals" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists the source portals available through this tool, with metadata such as jurisdiction and language. A jurisdiction may have multiple portals, including those maintained by regulators or ministries. When to use: When you need portals, legal websites, or sources for a known iso. Pass optional name, tag, or lang only to narrow that list. When not: Do not use this to read a document. For metadata use list_documents. For plain text use list_full_text. For structured Markdown use list_markdown. For a complete citation bundle use evidence_pull. For a temporary source-file URL use document_presigned_url. Do not call without an iso from list_jurisdictions. Skip if you already have this portal list for the same iso and filters. Parameters: iso (required) is a jurisdiction code from list_jurisdictions (e.g. BE, or US_CA for a state). name, tag, and lang are optional and combined with AND: name is a case-insensitive substring of the portal name, tag must match one portal tag exactly, lang must match the language name exactly (e.g. English). With limit, next_id parameters to paginate the results. Empty data means the iso is outside the subscription, or no portal with crawled links matched. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads current subscription and portal data. Portals with total_links of 0 are omitted. Returns: A list of portals matching the filters, with pagination information.
Annotations the server declares
- title: List portals
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_subcategoriesLists subcategories under one category (id, category_id, category_name, subcategory_id, subcategory_name).Reads text others write2 findings
Why it is in this group
- The name "list_subcategories" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Info
The description is 2,054 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...ubcategory_name":"Jurisdiction"}. Pass subcategory_id, not id, to list_law_types."
Description, as the server wrote it Lists subcategories under one category (id, category_id, category_name, subcategory_id, subcategory_name). This is the fourth taxonomy step. When to use: When you already have a category_id from list_categories. Examples: 'subcategories of category A.1.1', 'find the Jurisdiction subcategory'. Pass subcategory_name only to narrow by name. When not: Do not use this for law types; use list_law_types with the returned subcategory_id. For categories use list_categories. For domains or subdomains use list_domains or list_subdomains. For documents use list_documents or search. Do not call without a category_id from list_categories. Skip if you already have this subcategory list for the same category_id and subcategory_name. Parameters: category_id (required) is the category reference from list_categories. It must be exactly three dot-separated segments ending in a number, such as A.1.1, and is uppercased before the query. subcategory_name is an optional case-insensitive substring of the subcategory name; the parameter is subcategory_name, not subcategory. An unknown category_id returns Category not found. A category_id with the wrong shape is rejected before the query. limit defaults to 20 and is capped at 100. next_id is the numeric id from pagination.next_id, not subcategory_id. Empty data means that category has no matching subcategory. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads the current subcategory table. total_count is the full match count, not the page size. Returns: A page of {id, category_id, category_name, subcategory_id, subcategory_name}. Example: {"category_id":"A.1.1","category_name":"General Private International Law","subcategory_id":"A.1.1.1","subcategory_name":"Jurisdiction"}. Pass subcategory_id, not id, to list_law_types.
Annotations the server declares
- title: List subcategories
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
list_subdomainsLists subdomains under one domain (id, domain_id, subdomain_id, subdomain name).Reads text others write1 finding
Why it is in this group
- The name "list_subdomains" 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
Description, as the server wrote it Lists subdomains under one domain (id, domain_id, subdomain_id, subdomain name). This is the second taxonomy step. When to use: When you already have a domain_id from list_domains. Examples: 'subdomains under domain A', 'international law under Private law'. Pass subdomain only to narrow by name. When not: Do not call without a domain_id from list_domains; use list_domains first. For categories under a subdomain use list_categories with subdomain_id. For subcategories use list_subcategories. For law types use list_law_types. For documents use list_documents or search. Skip if you already have this subdomain list for the same domain_id and subdomain filter. Parameters: domain_id (required) is the domain reference from list_domains, such as A, and is uppercased before the query. It is not the numeric id. subdomain is an optional case-insensitive substring of the subdomain name. An unknown domain_id returns Domain not found. limit defaults to 20 and is capped at 100. next_id is the numeric id from pagination.next_id, not subdomain_id. Empty data means that domain has no matching subdomain. Behavior: Read-only SELECT; does not create, update, or delete records, files, or credentials. The only effect is a database read. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Not cached; each call reads the current subdomain table. total_count is the full match count, not the page size. Returns: A page of {id, domain_id, subdomain_id, subdomain}. Example: {"id":1,"domain_id":"A","subdomain_id":"A.1","subdomain":"Private International Law"}. Pass subdomain_id, not id, to list_categories.
Annotations the server declares
- title: List subdomains
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
searchSearches legal documents using OpenSearch to discover relevant documents based on the supplied search criteria.Reads private dataReads text others writeCan send data out2 findings
Why it is in this group
- The name "search" says "search": a word for reading.
What it can reach
- private-dataThe description's first sentence says "search" and "document": it reads documents.
- private-dataThe input names a file (input.file_data_only), and the tool reads: it reads files.
- 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.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...or delete documents, records, files, or credentials. The primary side effect is the computa..."
- Info
The description is 2,787 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...the search query, jurisdiction, supported filters, and subscription permissions."
Description, as the server wrote it Searches legal documents using OpenSearch to discover relevant documents based on the supplied search criteria. Returns matching documents and their available metadata. When to use: Use when you need to discover relevant legal documents through search, especially when looking for documents matching a search query rather than retrieving a structured list. Use this to explore legal content relevant to a topic, legal issue, phrase, or other supported search criteria. When not: Do not use this to browse with structured filters and cursor pagination; use list_documents. Do not use it to read a body or a file. For plain text use list_full_text (example: 'quote the operative wording'). For structured Markdown use list_markdown (example: 'keep the article headings'). For a complete citation bundle use evidence_pull (example: 'show me the source'). For a temporary source-file URL use document_presigned_url (example: 'download the PDF'). For one historical version use document_version. Do not call without a valid iso from list_jurisdictions. Avoid repeating an equivalent search when you already have sufficient results for the same query and filters. Parameters: iso (required) is the jurisdiction code from list_jurisdictions (e.g. BE or US_CA for a state). Provide the search query and any supported optional filters defined by the input schema. Optional filters may narrow the search by document metadata, dates, authority, source identifiers, tags, or other supported criteria. Use only parameters defined by the schema. The results depend on the query, jurisdiction, filters, and the user's subscription access. An empty data array means no matching documents were returned; it does not necessarily mean that the jurisdiction has no documents. Behavior: Read-only search operation. Does not create, update, or delete documents, records, files, or credentials. The primary side effect is the computational and resource cost of executing an OpenSearch query. Auth: tools/call requires OAuth 2.0 or a Bearer token (401 if missing, invalid, revoked, or no active subscription). Rate limits: none at the MCP layer; safe to retry. Search results are determined by the OpenSearch query implementation, relevance logic, supported filters, and available indexed data. Results may not include every document in the jurisdiction. Search results may differ from list_documents because the tools use different retrieval mechanisms. Repeating a search executes another query. Do not assume results are cached or that the search index contains every recently updated document. Returns: Matching legal documents and their available metadata, together with a result message. The returned documents depend on the search query, jurisdiction, supported filters, and subscription permissions.
Annotations the server declares
- title: Search
- 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.
576 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 4 outbound tools breaks every one.
- Reads private data8
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write18
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out4
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__lawstronaut__list_documents",
"mcp__lawstronaut__list_full_text",
"mcp__lawstronaut__list_markdown",
"mcp__lawstronaut__search"
]
}
}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 20, 2026, 18:41 UTC, unchanged since Nov 13, 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
- 2
- 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.lawstronaut.com/.well-known/oauth-protected-resource
- Authorization servers
- https://mcp.lawstronaut.com
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- yes
- Client ID metadata documents
- no
- WarnThe whole server
The endpoint served tools/list without credentials, and the list has 2 tools that change things (2 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://mcp.lawstronaut.com: tools/list returned 18 tools to an anonymous client, including horizon_scan (write), health_check (write)
Posture changes
- Nov 13, 2026
- Tools that change things, listed without credentials: 1 to 2
Findings
Matches of fixed checks, with the text they matched.
The description mentions "credential".
What this check looks forMatched text
description: "...know which iso codes are valid for this credential. When not: Skip if you already have thi..."
- Infolist_domains
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infolist_subdomains
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infolist_categories
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
The description is 2,054 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...ubcategory_name":"Jurisdiction"}. Pass subcategory_id, not id, to list_law_types."
Show 17 more findingsShow fewer
- Infolist_law_types
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infolist_portals
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infolist_documents
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infolist_documents
The description is 2,395 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...Returns: A list of documents matching the filters, with pagination information."
- Infolist_full_text
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Infolist_full_text
The description is 2,220 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...n text. Empty data means nothing matched, or the iso is outside the subscription."
- Infolist_markdown
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Infolist_markdown
The description is 2,290 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...rkdown. Empty data means nothing matched, or the iso is outside the subscription."
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Infodocument_version
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infohorizon_scan
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. The only effect is a database read. Au..."
- Infoevidence_pull
The description mentions "credentials".
What this check looks forMatched text
description: "...e, update, or delete records, files, or credentials. Auth: tools/call requires OAuth 2.0 or..."
- Infoevidence_pull
The description is 2,303 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "..._template, and, when requested, content_format plus content, and source_file_url."
- Infosearch
The description mentions "credentials".
What this check looks forMatched text
description: "...or delete documents, records, files, or credentials. The primary side effect is the computa..."
- Infosearch
The description is 2,787 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...the search query, jurisdiction, supported filters, and subscription permissions."
History
Changes to its tool definitions and to its remote endpoint, newest first.
PostureRemote posture changed. Tools that change things, listed without credentials: 1 -> 2.View diff
- Tools that change things, listed without credentials: 1 to 2
Tools6 added, 4 removed, 12 changed.+6 added−4 removed~12 changedView diff
- Added:document_presigned_urlOnly reads
- Added:document_versionOnly reads
- Added:health_checkChanges things
- Added:list_full_textOnly reads
- Added:list_markdownOnly reads
- Added:searchOnly reads
- Removed:get_document_text
- Removed:get_document_with_version
- Removed:get_markdown
- Removed:get_source_url
- Changed:evidence_pullOnly readsdescriptioninputSchema
- Changed:horizon_scanChanges thingsdescriptioninputSchema
- Changed:list_authority_typesOnly readsdescriptioninputSchema
- Changed:list_categoriesOnly readsdescriptioninputSchema
- Changed:list_documentsOnly readsdescriptioninputSchema
- Changed:list_domainsOnly readsdescriptioninputSchema
- Changed:list_issuing_authoritiesOnly readsdescriptioninputSchema
- Changed:list_jurisdictionsOnly readsdescriptioninputSchema
- Changed:list_law_typesOnly readsdescriptioninputSchema
- Changed:list_portalsOnly readsdescriptioninputSchema
- Changed:list_subcategoriesOnly readsdescriptioninputSchema
- Changed:list_subdomainsOnly readsdescriptioninputSchema
Install
Pinned to version 2.0.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http lawstronaut https://mcp.lawstronaut.comclaude mcp add --transport http lawstronaut https://mcp.lawstronaut.comRun 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__lawstronaut__list_jurisdictions",
"mcp__lawstronaut__list_domains",
"mcp__lawstronaut__list_subdomains",
"mcp__lawstronaut__list_categories",
"mcp__lawstronaut__list_subcategories",
"mcp__lawstronaut__list_law_types",
"mcp__lawstronaut__list_portals",
"mcp__lawstronaut__list_authority_types",
"mcp__lawstronaut__list_issuing_authorities",
"mcp__lawstronaut__document_presigned_url",
"mcp__lawstronaut__document_version",
"mcp__lawstronaut__evidence_pull"
],
"ask": [
"mcp__lawstronaut__list_documents",
"mcp__lawstronaut__list_full_text",
"mcp__lawstronaut__list_markdown",
"mcp__lawstronaut__horizon_scan",
"mcp__lawstronaut__search",
"mcp__lawstronaut__health_check"
],
"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 lawstronaut. Review it before you use it.
Capability tags behind 18 rules
- mcp__lawstronaut__list_jurisdictions private-data, untrusted-input
- mcp__lawstronaut__list_domains untrusted-input
- mcp__lawstronaut__list_subdomains untrusted-input
- mcp__lawstronaut__list_categories untrusted-input
- mcp__lawstronaut__list_subcategories untrusted-input
- mcp__lawstronaut__list_law_types untrusted-input
- mcp__lawstronaut__list_portals untrusted-input
- mcp__lawstronaut__list_authority_types untrusted-input
- mcp__lawstronaut__list_issuing_authorities untrusted-input
- mcp__lawstronaut__list_documents private-data, untrusted-input, outbound
- mcp__lawstronaut__list_full_text private-data, untrusted-input, outbound
- mcp__lawstronaut__list_markdown private-data, untrusted-input, outbound
- mcp__lawstronaut__document_presigned_url private-data, untrusted-input
- mcp__lawstronaut__document_version private-data, untrusted-input
- mcp__lawstronaut__horizon_scan untrusted-input
- mcp__lawstronaut__evidence_pull private-data, untrusted-input
- mcp__lawstronaut__search private-data, untrusted-input, outbound
- mcp__lawstronaut__health_check 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.lawstronaut/lawstronaut-mcp/verify?hash=a71f8b3f7d98d2cf193a0bb28701983db8fd8d74039185c7d713bc870c8fad9d" | jq -e '.changed == false'Pinned hash a71f8b3f7d98, from Nov 9, 2026, 16:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/com.lawstronaut/lawstronaut-mcp/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/com.lawstronaut/lawstronaut-mcp)Source: MCP Registry · repository. Synced Nov 9, 2026, 16:17 UTC.
Lawstronaut MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Lawstronaut MCP server safe?
- No scan can promise that. What its definitions show: 18 tools, 2 of which can change, send or delete something, and 23 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 Lawstronaut MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the Lawstronaut 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 Lawstronaut MCP server have?
- 18 tools: 2 write, 16 read. horizon_scan, health_check, list_jurisdictions, list_domains, list_subdomains, list_categories, list_subcategories, list_law_types and 10 more.
Does the Lawstronaut MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Lawstronaut MCP server send data to outside services?
- 4 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): list_documents, list_full_text, list_markdown and search.
Has the Lawstronaut MCP server changed its tools?
- Yes, once since Anlyon Radar started tracking it, most recently on Nov 9, 2026. Every change is on this page and in its Atom feed.