Create, validate and audit llms.txt, incl. the Lighthouse Agentic Browsing check.
Scored across 4 tools
The four tools split cleanly along input type (pasted text vs. URL vs. generation) and the descriptions explicitly cross-reference each other to steer selection. The only real overlap is between check_lighthouse_agentic_browsing and validate_llms_txt, which both analyze pasted text and both return a Lighthouse result, though the narrow-vs-full scope difference is spelled out.
All names are verb-first snake_case (check_*, check_*, generate_*, validate_*), which is predictable and readable. The mild deviation is using near-synonymous verbs 'check' and 'validate' for closely related analysis operations, which slightly blurs the pattern.
Four tools is well-scoped for a focused llms.txt generator/validator server, and each maps to a distinct input mode. It is a touch thin, with check_published_llms_txt partly duplicating checks available elsewhere rather than adding a wholly separate capability.
The surface covers the main lifecycle: generate a draft, validate pasted text, check a specific Lighthouse rule, and inspect a published file. Missing are any 'fix/rewrite' or emit-to-file operation and multi-domain handling, but these are minor gaps an agent can work around.
Use this when the user only wants to know whether a given llms.txt text passes the Lighthouse 13.3 "Agentic Browsing" audit (llms-txt). Applies exactly its three rules: an H1 exists, at least one Markdown link, more than 50 characters. On failure it returns Lighthouse's original message. Works only on the text provided and does not fetch anything. For a full format check use validate_llms_txt.
| Name | Required | Description | Default |
|---|---|---|---|
| content | Yes | The text of the llms.txt file. |
| Name | Required | Description |
|---|---|---|
| note | No | |
| checks | Yes | |
| passed | Yes | |
| reference | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover read-only, non-destructive, idempotent and closed-world traits, yet the description adds real value: the exact three rules applied, that failure returns Lighthouse's original message, and that it only inspects the supplied text and fetches nothing. Mostly complete, though success-side behavior is left to the output schema.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Four tight sentences, front-loaded with the trigger condition, then rules, then failure behavior, then the sibling alternative. No repetition or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With annotations, a fully documented single-parameter schema, and an output schema covering return values, the description supplies everything an agent needs — scope limits, rule set, and the alternative tool — without redundancy.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% for the single 'content' parameter, so the schema already defines it as the llms.txt text. The description adds only marginal framing ('uses only the text provided') and no format or size guidance (e.g. the 200k maxLength).
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb+resource: check whether llms.txt text passes the Lighthouse 13.3 'Agentic Browsing' audit. It also names the exact audit id (llms-txt) and its three rules, clearly separating it from validate_llms_txt.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicit when-to-use ('when the user only wants to know whether a given llms.txt text passes...') and an explicit alternative ('For a full format check use validate_llms_txt'), so the routing decision is unambiguous.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Use this when the user asks whether a website already has an llms.txt, whether it is served correctly, or why PageSpeed Insights / Lighthouse flags it. Requests exactly one URL, https:///llms.txt, and reports HTTP status, content type, redirects, the file content, the format validation and the Lighthouse result. Does not crawl other pages and does not change anything on the site. Messages are in German.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | Domain or any URL of the website, e.g. "example.com". |
| Name | Required | Description |
|---|---|---|
| url | Yes | |
| note | No | |
| found | Yes | |
| status | Yes | |
| content | No | |
| message | No | |
| redirects | No | |
| lighthouse | No | Result of the three Lighthouse 13.3 "llms-txt" audit rules. |
| validation | No | Format check against the llmstxt.org proposal. |
| contentType | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover the safety profile (readOnly, idempotent, non-destructive), and the description still adds substantial context: exactly one request to https://<domain>/llms.txt, the reported fields (status, content type, redirects, content, validation, Lighthouse result), no crawling, and the non-obvious behavioral trait that messages are returned in German.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loads the when-to-use triggers, then the request behavior, then the negative constraints. Dense but every sentence carries information; the trailing 'Messages are in German' is short and materially useful rather than filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
An output schema exists, yet the description still enumerates the returned fields, so the agent knows what it will get back. Combined with the safety notes and the German-language disclosure, nothing needed to select or invoke the tool correctly is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% and there is a single parameter, so the baseline is 3. The description goes beyond the schema by specifying what the tool does with the input: a domain or any site URL is normalized into a single GET of https://<domain>/llms.txt, which tells the agent the input need not already be the file URL.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb+resource (check a website's *published* llms.txt) and scopes it precisely: it fetches the live file rather than generating or validating a local one. That scope distinguishes it from generate_llms_txt and validate_llms_txt without opening their schemas.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives explicit trigger conditions ('when the user asks whether a website already has an llms.txt, whether it is served correctly, or why Lighthouse flags it') and an exclusion ('does not crawl other pages'). It stops short of naming the sibling tools as alternatives for the generate/validate cases, so it is clear but not fully routed.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Use this when the user wants a new llms.txt file for a public website they name. Reads up to 15 public HTML pages of that one domain (start page, XML sitemap, internal links) and returns a ready-to-publish llms.txt draft in Markdown, plus a format validation score, the Lighthouse "Agentic Browsing" result, the status of any llms.txt already published on the site, and the URLs that were left out with the reason. Deterministic: no language model is involved, every title and description is taken from the pages themselves. Limitations: honors robots.txt and stops if the start page is disallowed; does not render JavaScript, does not log in, and does not follow links to other domains; one request can take up to about a minute. Do not use this to fetch arbitrary page content or for sites the user only mentions in passing. Result messages are in German.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | The website, as a bare domain or full URL, e.g. "example.com" or "https://www.example.com". | |
| maxPages | No | Upper limit of pages to read. Default 15. |
| Name | Required | Description |
|---|---|---|
| note | No | |
| excluded | Yes | |
| finalUrl | Yes | |
| llmsText | Yes | The complete llms.txt draft in Markdown. |
| maxPages | No | |
| siteName | No | |
| warnings | Yes | |
| pageCount | Yes | |
| lighthouse | Yes | Result of the three Lighthouse 13.3 "llms-txt" audit rules. |
| validation | Yes | Format check against the llmstxt.org proposal. |
| limitReached | Yes | true when the page limit stopped the crawl. |
| publishedFile | No | Status of the llms.txt currently published on the site. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations cover the safety profile (readOnly, idempotent, non-destructive, openWorld), and the description goes well beyond them: deterministic/no LLM, 15-page cap, robots.txt honoring, no JS rendering, no login, no cross-domain following, ~1 minute latency, and German result messages. These are exactly the operational facts an agent cannot get from the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with the trigger condition, then outputs, then constraints. Every sentence carries distinct information — the deterministic note, the limitations list, and the anti-use cases all earn their place for a tool with this much behavior.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a one-shot generation tool with an output schema already present, the description supplies everything else an agent needs: scope, page budget, runtime, language of results, and failure conditions (robots.txt denial). Nothing material is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so the baseline is 3; the description adds value by clarifying that only the named domain is read, that the page budget defaults to 15, and that it is spent on the start page, XML sitemap, and internal links. It does not restate the bare-domain-vs-URL syntax beyond what the schema already documents.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The definition names a specific verb and artifact (generate an llms.txt draft) scoped to 'a public website they name', and it distinguishes itself from the check_* / validate_* siblings by stating it produces a new file rather than inspecting an existing one. The outputs are enumerated explicitly (draft, validation score, Lighthouse result, published-file status, skipped URLs).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It opens with the triggering condition ('when the user wants a new llms.txt file for a public website they name') and closes with explicit exclusions ('Do not use this to fetch arbitrary page content or for sites the user only mentions in passing'). The boundary against the sibling check/validate tools is implicit but clear from the 'new file' framing.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Use this when the user pastes the text of an llms.txt file and wants it checked or improved. Checks the text against the llmstxt.org format proposal: H1 title, summary blockquote, H2 sections, absolute Markdown links, duplicates, links to other domains, file size and update date. Returns a 0-100 score, findings with severity and line number, and the Lighthouse "Agentic Browsing" result. Works only on the text provided; it does not fetch anything from the internet. Finding messages are in German.
| Name | Required | Description | Default |
|---|---|---|---|
| content | Yes | The full text of the llms.txt file. | |
| expectedHostname | No | Optional hostname of the website, e.g. "www.example.com". Enables the warning for links to other domains. |
| Name | Required | Description |
|---|---|---|
| note | No | |
| lighthouse | Yes | Result of the three Lighthouse 13.3 "llms-txt" audit rules. |
| validation | Yes | Format check against the llmstxt.org proposal. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover safety (read-only, idempotent, closed-world), and the description adds substantive behavior: a 0-100 score, findings with severity and line numbers, a Lighthouse result, and notably that finding messages are in German. This last detail is non-obvious and materially affects how an agent presents results.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with the usage trigger, then the check list, return payload, and the no-fetch/language caveats. Four sentences with little waste, though the check enumeration is dense and could be trimmed.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With an output schema present, the description need not document returns, yet it summarizes the payload shape and flags the German-language output and offline constraint. Nothing an agent needs to call or interpret this tool is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% with only two params, so the schema carries the semantics. The description hints at expectedHostname's effect via the 'links to other domains' check but adds no format or syntax beyond the schema's example.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource (validate llms.txt content) and enumerates exactly what is checked (H1, blockquote, H2 sections, links, duplicates, size, update date). The 'works only on the text provided; does not fetch' line cleanly distinguishes it from the fetching sibling check_published_llms_txt.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives an explicit trigger ('when the user pastes the text... and wants it checked or improved') and a boundary (no fetching), which separates it from the URL-based sibling. It stops short of naming the alternative tool directly, so the routing is clear but not fully explicit.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
check_lighthouse_agentic_browsing2 fields changedInput schema / properties / content / descriptionPrevious value: -"Der Inhalt der llms.txt."New value: +"The text of the llms.txt file."Output schema / (root)Previous value: -nullNew value: +{
+ "properties": {
+ "checks": {
+ "items": {
+ "properties": {
+ "failureMessage": {
+ "type": "string"
+ },
+ "id": {
+ "enum": [
+ "h1",
+ "links",
+ "length"
+ ],
+ "type": "string"
+ },
+ "label": {
+ "type": "string"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "id",
+ "passed"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "note": {
+ "type": "string"
+ },
+ "passed": {
+ "type": "boolean"
+ },
+ "reference": {
+ "type": "string"
+ }
+ },
+ "required": [
+ "passed",
+ "checks"
+ ],
+ "type": "object"
+}check_published_llms_txt2 fields changedInput schema / properties / url / descriptionPrevious value: -"Domain oder beliebige URL der Website, z. B. \"beispiel.de\"."New value: +"Domain or any URL of the website, e.g. \"example.com\"."Output schema / (root)Previous value: -nullNew value: +{
+ "properties": {
+ "content": {
+ "type": "string"
+ },
+ "contentType": {
+ "type": "string"
+ },
+ "found": {
+ "type": "boolean"
+ },
+ "lighthouse": {
+ "description": "Result of the three Lighthouse 13.3 \"llms-txt\" audit rules.",
+ "properties": {
+ "checks": {
+ "items": {
+ "properties": {
+ "failureMessage": {
+ "type": "string"
+ },
+ "id": {
+ "enum": [
+ "h1",
+ "links",
+ "length"
+ ],
+ "type": "string"
+ },
+ "label": {
+ "type": "string"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "id",
+ "passed"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "passed",
+ "checks"
+ ],
+ "type": "object"
+ },
+ "message": {
+ "type": "string"
+ },
+ "note": {
+ "type": "string"
+ },
+ "redirects": {
+ "items": {
+ "type": "string"
+ },
+ "type": "array"
+ },
+ "status": {
+ "type": "integer"
+ },
+ "url": {
+ "type": "string"
+ },
+ "validation": {
+ "description": "Format check against the llmstxt.org proposal.",
+ "properties": {
+ "byteSize": {
+ "type": "number"
+ },
+ "issues": {
+ "items": {
+ "properties": {
+ "code": {
+ "type": "string"
+ },
+ "line": {
+ "type": "number"
+ },
+ "message": {
+ "type": "string"
+ },
+ "severity": {
+ "enum": [
+ "error",
+ "warning",
+ "info"
+ ],
+ "type": "string"
+ }
+ },
+ "required": [
+ "code",
+ "severity",
+ "message"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "links": {
+ "type": "number"
+ },
+ "score": {
+ "maximum": 100,
+ "minimum": 0,
+ "type": "number"
+ },
+ "valid": {
+ "description": "true when no error-level finding exists.",
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "valid",
+ "score",
+ "issues"
+ ],
+ "type": "object"
+ }
+ },
+ "required": [
+ "url",
+ "found",
+ "status"
+ ],
+ "type": "object"
+}generate_llms_txt3 fields changedInput schema / properties / maxPages / descriptionPrevious value: -"Obergrenze der Seiten, Standard 15."New value: +"Upper limit of pages to read. Default 15."Input schema / properties / url / descriptionPrevious value: -"Website, z. B. \"beispiel.de\" oder \"https://www.beispiel.de\"."New value: +"The website, as a bare domain or full URL, e.g. \"example.com\" or \"https://www.example.com\"."Output schema / (root)Previous value: -nullNew value: +{
+ "properties": {
+ "excluded": {
+ "items": {
+ "properties": {
+ "reason": {
+ "type": "string"
+ },
+ "url": {
+ "type": "string"
+ }
+ },
+ "required": [
+ "url",
+ "reason"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "finalUrl": {
+ "type": "string"
+ },
+ "lighthouse": {
+ "description": "Result of the three Lighthouse 13.3 \"llms-txt\" audit rules.",
+ "properties": {
+ "checks": {
+ "items": {
+ "properties": {
+ "failureMessage": {
+ "type": "string"
+ },
+ "id": {
+ "enum": [
+ "h1",
+ "links",
+ "length"
+ ],
+ "type": "string"
+ },
+ "label": {
+ "type": "string"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "id",
+ "passed"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "passed",
+ "checks"
+ ],
+ "type": "object"
+ },
+ "limitReached": {
+ "description": "true when the page limit stopped the crawl.",
+ "type": "boolean"
+ },
+ "llmsText": {
+ "description": "The complete llms.txt draft in Markdown.",
+ "type": "string"
+ },
+ "maxPages": {
+ "type": "integer"
+ },
+ "note": {
+ "type": "string"
+ },
+ "pageCount": {
+ "type": "integer"
+ },
+ "publishedFile": {
+ "description": "Status of the llms.txt currently published on the site.",
+ "type": "object"
+ },
+ "siteName": {
+ "type": "string"
+ },
+ "validation": {
+ "description": "Format check against the llmstxt.org proposal.",
+ "properties": {
+ "byteSize": {
+ "type": "number"
+ },
+ "issues": {
+ "items": {
+ "properties": {
+ "code": {
+ "type": "string"
+ },
+ "line": {
+ "type": "number"
+ },
+ "message": {
+ "type": "string"
+ },
+ "severity": {
+ "enum": [
+ "error",
+ "warning",
+ "info"
+ ],
+ "type": "string"
+ }
+ },
+ "required": [
+ "code",
+ "severity",
+ "message"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "links": {
+ "type": "number"
+ },
+ "score": {
+ "maximum": 100,
+ "minimum": 0,
+ "type": "number"
+ },
+ "valid": {
+ "description": "true when no error-level finding exists.",
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "valid",
+ "score",
+ "issues"
+ ],
+ "type": "object"
+ },
+ "warnings": {
+ "items": {
+ "type": "string"
+ },
+ "type": "array"
+ }
+ },
+ "required": [
+ "llmsText",
+ "finalUrl",
+ "pageCount",
+ "limitReached",
+ "validation",
+ "lighthouse",
+ "excluded",
+ "warnings"
+ ],
+ "type": "object"
+}validate_llms_txt3 fields changedInput schema / properties / content / descriptionPrevious value: -"Der vollständige Inhalt der llms.txt."New value: +"The full text of the llms.txt file."Input schema / properties / expectedHostname / descriptionPrevious value: -"Hostname der eigenen Website; aktiviert die Warnung bei Links auf fremde Domains."New value: +"Optional hostname of the website, e.g. \"www.example.com\". Enables the warning for links to other domains."Output schema / (root)Previous value: -nullNew value: +{
+ "properties": {
+ "lighthouse": {
+ "description": "Result of the three Lighthouse 13.3 \"llms-txt\" audit rules.",
+ "properties": {
+ "checks": {
+ "items": {
+ "properties": {
+ "failureMessage": {
+ "type": "string"
+ },
+ "id": {
+ "enum": [
+ "h1",
+ "links",
+ "length"
+ ],
+ "type": "string"
+ },
+ "label": {
+ "type": "string"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "id",
+ "passed"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "passed": {
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "passed",
+ "checks"
+ ],
+ "type": "object"
+ },
+ "note": {
+ "type": "string"
+ },
+ "validation": {
+ "description": "Format check against the llmstxt.org proposal.",
+ "properties": {
+ "byteSize": {
+ "type": "number"
+ },
+ "issues": {
+ "items": {
+ "properties": {
+ "code": {
+ "type": "string"
+ },
+ "line": {
+ "type": "number"
+ },
+ "message": {
+ "type": "string"
+ },
+ "severity": {
+ "enum": [
+ "error",
+ "warning",
+ "info"
+ ],
+ "type": "string"
+ }
+ },
+ "required": [
+ "code",
+ "severity",
+ "message"
+ ],
+ "type": "object"
+ },
+ "type": "array"
+ },
+ "links": {
+ "type": "number"
+ },
+ "score": {
+ "maximum": 100,
+ "minimum": 0,
+ "type": "number"
+ },
+ "valid": {
+ "description": "true when no error-level finding exists.",
+ "type": "boolean"
+ }
+ },
+ "required": [
+ "valid",
+ "score",
+ "issues"
+ ],
+ "type": "object"
+ }
+ },
+ "required": [
+ "validation",
+ "lighthouse"
+ ],
+ "type": "object"
+}check_lighthouse_agentic_browsingcheck_published_llms_txtgenerate_llms_txtvalidate_llms_txtAudit a site's visibility to ChatGPT, Perplexity, Claude and AI agents. Generates llms.txt.
Validate a site's llms.txt against the spec: fetches every linked URL, scores it 0-100.
Audit a page for search and AI answer engines; generate robots.txt, sitemap, head, llms.txt.
AI-search-visibility tools: check crawler access, generate llms.txt, lint robots.txt, census data.
Add one secure layer between your agents and this server.
