2pr.io writes, schedules and publishes LinkedIn posts in the user's own voice, and helps them reply to comments. It also writes posts for X, Telegram and Threads.
How to work with these tools:
- Call profile_list first. Almost every tool needs a profile_id. If a publish, schedule or reply fails because LinkedIn is not connected, tell the user to connect LinkedIn in 2pr.io at https://2pr.io/app.
- To write a post, call draft_create. It returns a url: always give the user that link so they can review and edit the draft in 2pr.io. Use draft_refine to change an existing draft instead of creating a new one.
- post_publish posts to LinkedIn immediately and publicly. Show the user the final text and get a clear yes before calling it. When the user gives a time, use post_schedule instead. Pass draftpost_id to either one to post a draft as it is.
- post_delete only cancels a scheduled post. It cannot remove a post that is already live on LinkedIn.
- LinkedIn is the default for everything. When the user asks for X, Telegram or Threads, write the post with draft_create and format set to "x", "telegram" or "threads" (draft_openFormat adds a network to an existing draft). Publish it with the x_* or telegram_* tools, only for a network the user named. Threads posts are drafts only.
- To add an image, get a hosted url first (media_upload, media_uploadFromUrl or media_generateImage), then call draft_attachImages.
- When the user says how they like to write ("shorter", "no emojis", "never use the word X"), save it with persona_appendDirective so every future draft follows it.
- To reply to comments: engage_getInbox, then engage_draftReplies, then engage_postReply only after the user picks or approves a reply.
- If a tool result has isError set, read the message: it says what was wrong with the call. Fix the arguments and try again, or tell the user what is needed.
Discoverability
2/6 passed2 of 6 passed33/100
Client listings and public setup information.
mcp.json or server-card.json
Either mcp.json or an MCP Server Card declaring this endpoint passes.
The manifest declares this MCP server's exact endpoint.
Tool definitions, annotations, and client compatibility.
All tools have descriptions
All tools have non-empty descriptions
Skills over MCP resources are readable
Server does not advertise the Skills over MCP extension
All tools have OpenAI behavior hints
47 behavior annotation issue(s) found
How to fix: On each tool, declare annotations.readOnlyHint, annotations.destructiveHint, and annotations.openWorldHint as booleans. The policy review separately assesses whether their values match the tool's actual behavior.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
List all profiles the caller has access to. Call this first to discover available profile_id values — most other tools require a profile_id. Also reports whether X is connected for each profile, so you can tell before attempting to publish there.
The description explains the capability, when to use it, and what to expect.
Inputs100/100
This tool declares no inputs, so argument descriptions are not needed.
No input fields.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "profile_list",
"description": "List all profiles the caller has access to. Call this first to discover available profile_id values — most other tools require a profile_id. Also reports whether X is connected for each profile, so you can tell before attempting to publish there.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {},
"additionalProperties": false
},
"annotations": {
"title": "List Profiles",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Create a new draft post from the user input text and render it as three variants of the SAME angle: `regular_body` (standard medium-form), `extended_body` (longer, more depth), and `original_body` (the raw frontier-model baseline: natural paragraphs, none of our house style layered on). The three variants are NOT different hook angles. For angle diversity (different framings of the same topic), call this tool multiple times with different `user_input_text` framings. `user_input_text` says only what the post is about and how to write it. Never ask it for an image, infographic or visual idea: the writer returns publishable post text and nothing else, so the idea would end up inside the post. When the user wants a visual with the post, create the draft first, then make the picture with media_generateImage or media_generateAnimation and attach it with draft_attachImages. To ground the post in a web page, article, YouTube video, LinkedIn post, or Reddit thread, pass its link as `source_url` and put the angle/instructions in `user_input_text` (e.g. "a contrarian take on this"); the server fetches and extracts the page and folds it into the generation. The source pictures (a LinkedIn post images, an article lead image) come back as `source_page_image_urls` and are shown to you; they are not attached to the draft. When the user then wants a picture for the post, pass them to media_generateImage as `reference_image_urls` so the new image builds on the original instead of being invented from the text. To draft FROM a file or voice note the user attached (its 2pr storage URL is given in the conversation context), pass that URL as `source_file_url`: documents (PDF/DOCX/PPTX/CSV/Markdown/TXT) are read as the source and audio/video (voice notes, MP4) are transcribed, with `user_input_text` the optional instruction alongside it. There are two ways to pass an image the user attached, and picking the wrong one is the most common way to ruin a draft. Use `source_image_urls` ONLY when the post should be ABOUT what is in the picture: each image is vision-described into the generation, which makes the body explain the image and state that the post is based on a file the user shared. Use `attach_image_urls` when the picture is just the visual and the user already supplied the words (a meme caption, a screenshot they are reacting to): the image is attached to the draft but kept out of the generation, so the post does not narrate what the reader can already see. Either way the image ships on the draft, so neither needs a separate draft_attachImages call. When the user hands over FINISHED post copy (written in this conversation or another tool) and wants it kept as is, set `verbatim: true`: the text in `user_input_text` is stored unchanged as all three bodies and no generation runs, so nothing is rewritten. Provide at least one of `user_input_text`, `source_url`, `source_file_url`, or `source_image_urls`. Generation runs synchronously so the returned `url` opens an already-populated editor. Takes ~10-20s (longer with `source_url`). If generation fails the draft is still created with empty bodies and the caller can retry from the editor. For a long-form LinkedIn ARTICLE or NEWSLETTER EDITION (not a feed post), set `format: "article"` (optionally `length`): the draft holds one article in markdown (a `# Title` line, `##` section headings) instead of three post variants, written from `user_input_text` and optionally `source_url` and/or `source_file_url` (a report, deck or voice note the user attached to build the article from); an image the user attached goes in `attach_image_urls` and becomes the article cover. It takes 1-2 minutes. LinkedIn does not let apps publish articles, so the draft editor at `url` has copy buttons that paste it into the LinkedIn article editor with formatting intact; do not try to publish or schedule an article. When the user wants a post for X, Telegram or Threads and has no draft for it yet, set `format` to "x", "telegram" or "threads": the post is written for that network straight from `user_input_text` (their words and any asks, such as "translate all of it", go in as they wrote them) and optionally `source_url`, with no LinkedIn post made first; `verbatim` and `attach_image_urls` work as usual. To add another network to a draft that already exists, use draft_openFormat instead.
The description explains the capability, when to use it, and what to expect.
Inputs67/100
6 of 9 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
user_input_text
Not provided
Missing description. Explain what this value means and how to supply it.
source_url
Optional public http(s) link to ground the post in (article, YouTube, LinkedIn post, Reddit thread). Its extracted content is folded into the generation alongside user_input_text.
Clear enough to supply this argument.
source_file_url
Not provided
Missing description. Explain what this value means and how to supply it.
source_image_urls
2pr storage URLs of attached images to draft FROM: each is vision-described into the generation and auto-attached to the new draft. Use the hosted URLs given in the conversation context for user attachments. Use this ONLY when the post should be ABOUT what is in the picture: it routes generation down the write-about-this-source path, which describes the image in the body and states that the post is based on a file the user shared. For a picture that simply ships with the post, use attach_image_urls instead.
Clear enough to supply this argument.
attach_image_urls
2pr storage URLs of images to ATTACH to the new draft WITHOUT feeding them into generation. Use this whenever the picture is the visual and the words come from the user: a meme the user wrote the caption for, a screenshot they are reacting to, a chart they are commenting on. The body is generated from user_input_text alone, so the post does not narrate what the reader can already see.
Clear enough to supply this argument.
verbatim
Set true when user_input_text is the finished post and must not be rewritten. The text is saved exactly as sent and no generation runs. Requires user_input_text; cannot be combined with source_url, source_file_url, or source_image_urls. attach_image_urls still works.
Clear enough to supply this argument.
format
Default "post" (LinkedIn). "article" writes a long-form LinkedIn article or newsletter edition instead of a feed post. "x", "telegram" or "threads" starts the idea as a post for that network, with no LinkedIn post.
Clear enough to supply this argument.
length
Article length only: short is about 800 words, standard about 1,500 (default), deep about 2,500.
Clear enough to supply this argument.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "draft_create",
"description": "Create a new draft post from the user input text and render it as three variants of the SAME angle: `regular_body` (standard medium-form), `extended_body` (longer, more depth), and `original_body` (the raw frontier-model baseline: natural paragraphs, none of our house style layered on). The three variants are NOT different hook angles. For angle diversity (different framings of the same topic), call this tool multiple times with different `user_input_text` framings. `user_input_text` says only what the post is about and how to write it. Never ask it for an image, infographic or visual idea: the writer returns publishable post text and nothing else, so the idea would end up inside the post. When the user wants a visual with the post, create the draft first, then make the picture with media_generateImage or media_generateAnimation and attach it with draft_attachImages. To ground the post in a web page, article, YouTube video, LinkedIn post, or Reddit thread, pass its link as `source_url` and put the angle/instructions in `user_input_text` (e.g. \"a contrarian take on this\"); the server fetches and extracts the page and folds it into the generation. The source pictures (a LinkedIn post images, an article lead image) come back as `source_page_image_urls` and are shown to you; they are not attached to the draft. When the user then wants a picture for the post, pass them to media_generateImage as `reference_image_urls` so the new image builds on the original instead of being invented from the text. To draft FROM a file or voice note the user attached (its 2pr storage URL is given in the conversation context), pass that URL as `source_file_url`: documents (PDF/DOCX/PPTX/CSV/Markdown/TXT) are read as the source and audio/video (voice notes, MP4) are transcribed, with `user_input_text` the optional instruction alongside it. There are two ways to pass an image the user attached, and picking the wrong one is the most common way to ruin a draft. Use `source_image_urls` ONLY when the post should be ABOUT what is in the picture: each image is vision-described into the generation, which makes the body explain the image and state that the post is based on a file the user shared. Use `attach_image_urls` when the picture is just the visual and the user already supplied the words (a meme caption, a screenshot they are reacting to): the image is attached to the draft but kept out of the generation, so the post does not narrate what the reader can already see. Either way the image ships on the draft, so neither needs a separate draft_attachImages call. When the user hands over FINISHED post copy (written in this conversation or another tool) and wants it kept as is, set `verbatim: true`: the text in `user_input_text` is stored unchanged as all three bodies and no generation runs, so nothing is rewritten. Provide at least one of `user_input_text`, `source_url`, `source_file_url`, or `source_image_urls`. Generation runs synchronously so the returned `url` opens an already-populated editor. Takes ~10-20s (longer with `source_url`). If generation fails the draft is still created with empty bodies and the caller can retry from the editor. For a long-form LinkedIn ARTICLE or NEWSLETTER EDITION (not a feed post), set `format: \"article\"` (optionally `length`): the draft holds one article in markdown (a `# Title` line, `##` section headings) instead of three post variants, written from `user_input_text` and optionally `source_url` and/or `source_file_url` (a report, deck or voice note the user attached to build the article from); an image the user attached goes in `attach_image_urls` and becomes the article cover. It takes 1-2 minutes. LinkedIn does not let apps publish articles, so the draft editor at `url` has copy buttons that paste it into the LinkedIn article editor with formatting intact; do not try to publish or schedule an article. When the user wants a post for X, Telegram or Threads and has no draft for it yet, set `format` to \"x\", \"telegram\" or \"threads\": the post is written for that network straight from `user_input_text` (their words and any asks, such as \"translate all of it\", go in as they wrote them) and optionally `source_url`, with no LinkedIn post made first; `verbatim` and `attach_image_urls` work as usual. To add another network to a draft that already exists, use draft_openFormat instead.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"format": {
"enum": [
"post",
"article",
"x",
"telegram",
"threads"
],
"type": "string",
"description": "Default \"post\" (LinkedIn). \"article\" writes a long-form LinkedIn article or newsletter edition instead of a feed post. \"x\", \"telegram\" or \"threads\" starts the idea as a post for that network, with no LinkedIn post."
},
"length": {
"enum": [
"short",
"standard",
"deep"
],
"type": "string",
"description": "Article length only: short is about 800 words, standard about 1,500 (default), deep about 2,500."
},
"verbatim": {
"type": "boolean",
"description": "Set true when user_input_text is the finished post and must not be rewritten. The text is saved exactly as sent and no generation runs. Requires user_input_text; cannot be combined with source_url, source_file_url, or source_image_urls. attach_image_urls still works."
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"source_url": {
"type": "string",
"format": "uri",
"description": "Optional public http(s) link to ground the post in (article, YouTube, LinkedIn post, Reddit thread). Its extracted content is folded into the generation alongside user_input_text."
},
"source_file_url": {
"type": "string",
"format": "uri"
},
"user_input_text": {
"type": "string",
"maxLength": 10000,
"minLength": 1
},
"attach_image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 9,
"minItems": 1,
"description": "2pr storage URLs of images to ATTACH to the new draft WITHOUT feeding them into generation. Use this whenever the picture is the visual and the words come from the user: a meme the user wrote the caption for, a screenshot they are reacting to, a chart they are commenting on. The body is generated from user_input_text alone, so the post does not narrate what the reader can already see."
},
"source_image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 4,
"minItems": 1,
"description": "2pr storage URLs of attached images to draft FROM: each is vision-described into the generation and auto-attached to the new draft. Use the hosted URLs given in the conversation context for user attachments. Use this ONLY when the post should be ABOUT what is in the picture: it routes generation down the write-about-this-source path, which describes the image in the body and states that the post is based on a file the user shared. For a picture that simply ships with the post, use attach_image_urls instead."
}
},
"additionalProperties": false
},
"annotations": {
"title": "Create Draft",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Read an existing draft by id. Returns the three variants of the same angle (`regular_body` standard, `extended_body` longer, `original_body` the raw frontier-model baseline in natural paragraphs), plus `first_comment`, any attached media (images/video/document) and the LinkedIn `poll` if the draft has one. On the chat surface the attached images are also shown to you as vision blocks alongside the result, so you can see and comment on their actual content. Use this before scheduling/publishing a draft when you need to inspect or copy its contents — or pass `draft_id` directly into post_publish/post_schedule and let them hydrate from the draft.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
draft_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "draft_get",
"description": "Read an existing draft by id. Returns the three variants of the same angle (`regular_body` standard, `extended_body` longer, `original_body` the raw frontier-model baseline in natural paragraphs), plus `first_comment`, any attached media (images/video/document) and the LinkedIn `poll` if the draft has one. On the chat surface the attached images are also shown to you as vision blocks alongside the result, so you can see and comment on their actual content. Use this before scheduling/publishing a draft when you need to inspect or copy its contents — or pass `draft_id` directly into post_publish/post_schedule and let them hydrate from the draft.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"draft_id"
],
"properties": {
"draft_id": {
"type": "string",
"format": "uuid"
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Draft",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description97/100
Replace a draft body WHOLESALE. Only the provided fields are written. Variants follow draft_create semantics: `regular_body` standard, `extended_body` longer, `original_body` the raw frontier-model baseline in natural paragraphs. Same angle, different treatment. Pass `variant` plus the MATCHING body field to rewrite a single variant and leave the other two untouched. WARNING: the body you send REPLACES what is stored, so any edit the user made in the editor since the draft was last read is overwritten. To attach or replace a LinkedIn poll on the draft, pass `poll` (it removes any images, video or document, since a poll cannot carry media); to drop the poll, set `remove_poll: true`. Returns the updated draft and a `url` deep link to the draft editor.
The description explains the capability, when to use it, and what to expect.
Inputs24/100
4 of 17 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
draft_id
Not provided
Missing description. Explain what this value means and how to supply it.
variant
Not provided
Missing description. Explain what this value means and how to supply it.
user_input_text
Not provided
Missing description. Explain what this value means and how to supply it.
regular_body
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
extended_body
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
original_body
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
first_comment
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment_image_url
Not provided
Missing description. Explain what this value means and how to supply it.
is_copied
Not provided
Missing description. Explain what this value means and how to supply it.
is_shared
Not provided
Missing description. Explain what this value means and how to supply it.
seo_title
Not provided
Missing description. Explain what this value means and how to supply it.
seo_description
Not provided
Missing description. Explain what this value means and how to supply it.
poll › question
Not provided
Missing description. Explain what this value means and how to supply it.
poll › options
Not provided
Missing description. Explain what this value means and how to supply it.
poll › duration
Not provided
Missing description. Explain what this value means and how to supply it.
remove_poll
Set true ONLY when the user asked to remove the poll from the draft.
Clear enough to supply this argument.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "draft_refine",
"description": "Replace a draft body WHOLESALE. Only the provided fields are written. Variants follow draft_create semantics: `regular_body` standard, `extended_body` longer, `original_body` the raw frontier-model baseline in natural paragraphs. Same angle, different treatment. Pass `variant` plus the MATCHING body field to rewrite a single variant and leave the other two untouched. WARNING: the body you send REPLACES what is stored, so any edit the user made in the editor since the draft was last read is overwritten. To attach or replace a LinkedIn poll on the draft, pass `poll` (it removes any images, video or document, since a poll cannot carry media); to drop the poll, set `remove_poll: true`. Returns the updated draft and a `url` deep link to the draft editor.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"draft_id"
],
"properties": {
"poll": {
"type": "object",
"required": [
"question",
"options",
"duration"
],
"properties": {
"options": {
"type": "array",
"items": {
"type": "string",
"maxLength": 30,
"minLength": 1
},
"maxItems": 4,
"minItems": 2
},
"duration": {
"enum": [
"ONE_DAY",
"THREE_DAYS",
"SEVEN_DAYS",
"FOURTEEN_DAYS"
],
"type": "string"
},
"question": {
"type": "string",
"maxLength": 140,
"minLength": 1
}
},
"description": "A LinkedIn poll to publish with the post. Send it ONLY when the user asked for a poll. It needs a question and answer options: at least two options, at most four, each a short phrase. `duration` is how long voting stays open. A poll cannot be combined with images, a video or a document, and it cannot be edited after it is published. The post text still goes out above the poll.",
"additionalProperties": false
},
"variant": {
"enum": [
"regular",
"extended",
"original"
],
"type": "string"
},
"draft_id": {
"type": "string",
"format": "uuid"
},
"is_copied": {
"type": "boolean"
},
"is_shared": {
"type": "boolean"
},
"seo_title": {
"anyOf": [
{
"type": "string",
"maxLength": 200
},
{
"type": "null"
}
]
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"remove_poll": {
"type": "boolean",
"description": "Set true ONLY when the user asked to remove the poll from the draft."
},
"regular_body": {
"type": [
"string",
"null"
],
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"extended_body": {
"type": [
"string",
"null"
],
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"first_comment": {
"type": [
"string",
"null"
]
},
"original_body": {
"type": [
"string",
"null"
],
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"seo_description": {
"anyOf": [
{
"type": "string",
"maxLength": 400
},
{
"type": "null"
}
]
},
"user_input_text": {
"type": [
"string",
"null"
]
},
"first_comment_image_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"annotations": {
"title": "Refine Draft",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Attach one or more images to a draft by their public URL so they show in the draft editor and go out with the post. Pass the 2pr-hosted URLs surfaced in the conversation context when the user attaches an image, or a URL returned by media_upload / media_uploadFromUrl. LinkedIn allows up to 9 images per post; attaching images replaces any existing video or document on the draft (a post can carry only one media kind). `alt_texts` is optional and index-aligned to `image_urls`. Returns the updated draft with its full `image_urls` list.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 4 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
draft_id
Not provided
Missing description. Explain what this value means and how to supply it.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
alt_texts
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "draft_attachImages",
"description": "Attach one or more images to a draft by their public URL so they show in the draft editor and go out with the post. Pass the 2pr-hosted URLs surfaced in the conversation context when the user attaches an image, or a URL returned by media_upload / media_uploadFromUrl. LinkedIn allows up to 9 images per post; attaching images replaces any existing video or document on the draft (a post can carry only one media kind). `alt_texts` is optional and index-aligned to `image_urls`. Returns the updated draft with its full `image_urls` list.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"draft_id",
"image_urls"
],
"properties": {
"draft_id": {
"type": "string",
"format": "uuid"
},
"alt_texts": {
"type": "array",
"items": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
]
}
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 9,
"minItems": 1
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Attach Images to Draft",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Attach a single video or PDF document to a draft by its public 2pr URL so it shows in the draft editor and goes out with the post. Use draft_attachImages for images instead. A LinkedIn post can carry only one media kind, so attaching a video or document replaces any images, video, or document already on the draft. `media_type` must match the file: `video` for mp4/mov, `document` for PDF. Pass a URL returned by media_upload / media_uploadFromUrl / media_commitUpload — their `kind` tells you which `media_type` to use. `title` is optional; for a document it becomes the LinkedIn document title. Returns the updated draft.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 5 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
draft_id
Not provided
Missing description. Explain what this value means and how to supply it.
media_type
Not provided
Missing description. Explain what this value means and how to supply it.
url
Not provided
Missing description. Explain what this value means and how to supply it.
title
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
{
"title": "Attach Media to Draft",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
Output schema0/100
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "draft_attachMedia",
"description": "Attach a single video or PDF document to a draft by its public 2pr URL so it shows in the draft editor and goes out with the post. Use draft_attachImages for images instead. A LinkedIn post can carry only one media kind, so attaching a video or document replaces any images, video, or document already on the draft. `media_type` must match the file: `video` for mp4/mov, `document` for PDF. Pass a URL returned by media_upload / media_uploadFromUrl / media_commitUpload — their `kind` tells you which `media_type` to use. `title` is optional; for a document it becomes the LinkedIn document title. Returns the updated draft.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"draft_id",
"media_type",
"url"
],
"properties": {
"url": {
"type": "string",
"format": "uri"
},
"title": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
]
},
"draft_id": {
"type": "string",
"format": "uuid"
},
"media_type": {
"enum": [
"video",
"document"
],
"type": "string"
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Attach Media to Draft",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description87/100
Search the viral LinkedIn post catalog by keyword, with optional filters (language, creator URLs, engagement thresholds, date range). Defaults to the last 12 months and reference-quality posts. Put the subject in search_query; every extra filter narrows the catalog, and results are already ranked by relevance weighted by engagement. Phrase search_query as the subject ("warm outreach"), not as the request ("show me posts about warm outreach"): every word has to appear in the post. The search broadens itself when a query is too narrow, so a thin result means the shorter query was tried too. There is no region or country filter. This catalog is LinkedIn only: for trending X (Twitter) posts, use inspiration_getTopics, whose feed includes viral X posts.
The description explains the capability, when to use it, and what to expect.
Inputs35/100
11 of 24 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
search_query
Not provided
Missing description. Explain what this value means and how to supply it.
languages
Not provided
Missing description. Explain what this value means and how to supply it.
user_url
Not provided
Missing description. Explain what this value means and how to supply it.
user_urls
Not provided
Missing description. Explain what this value means and how to supply it.
date_posted_start
Not provided
Missing description. Explain what this value means and how to supply it.
date_posted_end
Not provided
Missing description. Explain what this value means and how to supply it.
min_likes
Only when the user named a floor ("at least 1000 likes"). Omit otherwise.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
max_likes
Only when the user named a ceiling. Omit otherwise; never send 0.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
min_comments
Not provided
Missing description. Explain what this value means and how to supply it.
max_comments
Only when the user named a ceiling. Omit otherwise; never send 0.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
min_followers
Not provided
Missing description. Explain what this value means and how to supply it.
max_followers
Only when the user named a ceiling. Omit otherwise; never send 0.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
min_posts
Not provided
Missing description. Explain what this value means and how to supply it.
max_posts
Only when the user named a ceiling. Omit otherwise; never send 0.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
min_articles
Not provided
Missing description. Explain what this value means and how to supply it.
max_articles
Only when the user named a ceiling. Omit otherwise; never send 0.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
has_images
Set true to require an image. Omit for any post; false is ignored.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
has_videos
Set true to require a video. Omit for any post; false is ignored.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
has_carousel
Set true to require a carousel. Omit for any post; false is ignored.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
has_gif
Set true to require a GIF. Omit for any post; false is ignored.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
account_type
Not provided
Missing description. Explain what this value means and how to supply it.
sort_by
outliers keeps posts with 3x+ their creator's usual likes (median of the creator's recent posts) and ranks by that multiple. Only posts with 300+ likes qualify.
Clear enough to supply this argument.
limit
Not provided
Missing description. Explain what this value means and how to supply it.
offset
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare destructiveHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "viral_search",
"description": "Search the viral LinkedIn post catalog by keyword, with optional filters (language, creator URLs, engagement thresholds, date range). Defaults to the last 12 months and reference-quality posts. Put the subject in search_query; every extra filter narrows the catalog, and results are already ranked by relevance weighted by engagement. Phrase search_query as the subject (\"warm outreach\"), not as the request (\"show me posts about warm outreach\"): every word has to appear in the post. The search broadens itself when a query is too narrow, so a thin result means the shorter query was tried too. There is no region or country filter. This catalog is LinkedIn only: for trending X (Twitter) posts, use inspiration_getTopics, whose feed includes viral X posts.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {
"limit": {
"type": "integer",
"maximum": 50,
"minimum": 1
},
"offset": {
"type": "integer",
"minimum": 0
},
"has_gif": {
"type": "boolean",
"description": "Set true to require a GIF. Omit for any post; false is ignored."
},
"sort_by": {
"enum": [
"likes",
"date",
"relevance",
"outliers"
],
"type": "string",
"description": "outliers keeps posts with 3x+ their creator's usual likes (median of the creator's recent posts) and ranks by that multiple. Only posts with 300+ likes qualify."
},
"user_url": {
"type": "string"
},
"languages": {
"type": "string"
},
"max_likes": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a ceiling. Omit otherwise; never send 0."
},
"max_posts": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a ceiling. Omit otherwise; never send 0."
},
"min_likes": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a floor (\"at least 1000 likes\"). Omit otherwise."
},
"min_posts": {
"type": "integer",
"minimum": 0
},
"user_urls": {
"type": "array",
"items": {
"type": "string"
}
},
"has_images": {
"type": "boolean",
"description": "Set true to require an image. Omit for any post; false is ignored."
},
"has_videos": {
"type": "boolean",
"description": "Set true to require a video. Omit for any post; false is ignored."
},
"account_type": {
"enum": [
"person",
"company"
],
"type": "string"
},
"has_carousel": {
"type": "boolean",
"description": "Set true to require a carousel. Omit for any post; false is ignored."
},
"max_articles": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a ceiling. Omit otherwise; never send 0."
},
"max_comments": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a ceiling. Omit otherwise; never send 0."
},
"min_articles": {
"type": "integer",
"minimum": 0
},
"min_comments": {
"type": "integer",
"minimum": 0
},
"search_query": {
"type": "string",
"minLength": 1
},
"max_followers": {
"type": "integer",
"minimum": 0,
"description": "Only when the user named a ceiling. Omit otherwise; never send 0."
},
"min_followers": {
"type": "integer",
"minimum": 0
},
"date_posted_end": {
"type": "string"
},
"date_posted_start": {
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Search Viral Posts",
"readOnlyHint": true,
"openWorldHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description96/100
Read the user's current inspiration feed: the personalized post ideas 2pr.io generates from their sources: viral LinkedIn posts, viral X (Twitter) posts, Reddit, Hacker News, GitHub trending, newsletters and blogs they added, and their own posts and comments. Returns each idea's title, source, and a source label. Use when the user asks for post ideas, inspiration, or "what should I write about", and also when they ask for trending tweets or X posts to repurpose. Returns an empty list when the feed has not been generated yet.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "inspiration_getTopics",
"description": "Read the user's current inspiration feed: the personalized post ideas 2pr.io generates from their sources: viral LinkedIn posts, viral X (Twitter) posts, Reddit, Hacker News, GitHub trending, newsletters and blogs they added, and their own posts and comments. Returns each idea's title, source, and a source label. Use when the user asks for post ideas, inspiration, or \"what should I write about\", and also when they ask for trending tweets or X posts to repurpose. Returns an empty list when the feed has not been generated yet.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Inspiration Topics",
"readOnlyHint": true,
"idempotentHint": true
}
}
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "persona_get",
"description": "Get the persona / voice brief for a profile in markdown. Returns null content if no persona is set and no onboarding data exists to derive one.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Voice Profile",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description81/100
Append a voice/style directive (single bullet, at most 1000 characters) into the user-directives section of a profile's persona. Creates a baseline persona from onboarding if none exists. For anything longer, split it across several calls or use persona_replace.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
directive
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare openWorldHint based on the tool’s actual behavior.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description89/100
Read the content preferences for a profile: language, theme, audience, goal, persona, tone, reach-vs-authenticity priority, geography, emoji use, Engage reply instructions, and which inspiration sources, subreddits and feeds are in play. Use this before changing anything so you say what is actually set. Not the voice brief, which is persona_get.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "preferences_get",
"description": "Read the content preferences for a profile: language, theme, audience, goal, persona, tone, reach-vs-authenticity priority, geography, emoji use, Engage reply instructions, and which inspiration sources, subreddits and feeds are in play. Use this before changing anything so you say what is actually set. Not the voice brief, which is persona_get.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Preferences",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description93/100
Change content preferences for a profile. Send only the fields to change; pass null to clear one. It takes effect everywhere at once (drafting, the inspiration feed, Engage replies). Cannot set the voice brief (use persona_replace / persona_appendDirective) or any internal flag.
The description explains the capability, when to use it, and what to expect.
Inputs72/100
14 of 15 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
language
Language the posts are written in, as an English language name ("English", "Russian").
Clear enough to supply this argument.
content_primary_theme
Main subject the creator posts about. Changing it re-derives which subreddits we track.
Clear enough to supply this argument.
target_audience
Who the posts are written for.
Explain the meaning of this value, not just its name.
content_goal
What the creator wants the posts to achieve.
Clear enough to supply this argument.
content_persona
The role the creator writes as (founder, engineer, coach).
Clear enough to supply this argument.
content_tone
How the posts should sound (casual, direct, warm).
Clear enough to supply this argument.
content_priority
Balance between reach and authenticity. Canonical values: "Performance only", "Metrics first", "Balance", "Authenticity first", "Authenticity only".
Clear enough to supply this argument.
geographic_focus
Region the audience sits in.
Explain the meaning of this value, not just its name.
emoji_preference
How much emoji the posts should use.
Explain the meaning of this value, not just its name.
engage_custom_instructions
Standing instructions for the replies drafted in the Engage inbox.
Clear enough to supply this argument.
custom_subreddits
Comma-separated subreddit names to pull inspiration from, without the "r/" prefix. Empty falls back to theme-derived subreddits.
Clear enough to supply this argument.
inspiration_feeds
Comma-separated newsletter / blog feed URLs used by the inspiration feed.
Clear enough to supply this argument.
inspiration_recency_days
How far back the inspiration sources may look, in days.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description93/100
Publish a post to LinkedIn immediately. Persists a row in scheduled_posts with status=posted on success. Pass `draftpost_id` to hydrate text/image_urls/first_comment/video/document from the draft (defaults to the regular body — override with `body_variant`). Explicit fields take precedence over draft fields. `text` is optional when media is attached: a media-only call publishes with no caption. Set do_not_share:true to record the post as published without actually calling LinkedIn (used for tracking off-platform sharing). `first_comment_image_url` attaches an image (or GIF) to the first comment. To publish a LinkedIn poll, pass `poll` (question, options, duration) with the post text and no media; a draft that already carries a poll publishes it automatically.
The description explains the capability, when to use it, and what to expect.
Inputs7/100
1 of 15 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
text
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
author_urn
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment_image_url
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
body_variant
Not provided
Missing description. Explain what this value means and how to supply it.
do_not_share
Not provided
Missing description. Explain what this value means and how to supply it.
document_url
Not provided
Missing description. Explain what this value means and how to supply it.
document_title
Not provided
Missing description. Explain what this value means and how to supply it.
video_url
Not provided
Missing description. Explain what this value means and how to supply it.
poll › question
Not provided
Missing description. Explain what this value means and how to supply it.
poll › options
Not provided
Missing description. Explain what this value means and how to supply it.
poll › duration
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "post_publish",
"description": "Publish a post to LinkedIn immediately. Persists a row in scheduled_posts with status=posted on success. Pass `draftpost_id` to hydrate text/image_urls/first_comment/video/document from the draft (defaults to the regular body — override with `body_variant`). Explicit fields take precedence over draft fields. `text` is optional when media is attached: a media-only call publishes with no caption. Set do_not_share:true to record the post as published without actually calling LinkedIn (used for tracking off-platform sharing). `first_comment_image_url` attaches an image (or GIF) to the first comment. To publish a LinkedIn poll, pass `poll` (question, options, duration) with the post text and no media; a draft that already carries a poll publishes it automatically.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"poll": {
"type": "object",
"required": [
"question",
"options",
"duration"
],
"properties": {
"options": {
"type": "array",
"items": {
"type": "string",
"maxLength": 30,
"minLength": 1
},
"maxItems": 4,
"minItems": 2
},
"duration": {
"enum": [
"ONE_DAY",
"THREE_DAYS",
"SEVEN_DAYS",
"FOURTEEN_DAYS"
],
"type": "string"
},
"question": {
"type": "string",
"maxLength": 140,
"minLength": 1
}
},
"description": "A LinkedIn poll to publish with the post. Send it ONLY when the user asked for a poll. It needs a question and answer options: at least two options, at most four, each a short phrase. `duration` is how long voting stays open. A poll cannot be combined with images, a video or a document, and it cannot be edited after it is published. The post text still goes out above the poll.",
"additionalProperties": false
},
"text": {
"type": "string",
"minLength": 1,
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"video_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"author_urn": {
"anyOf": [
{
"type": "string",
"minLength": 1
},
{
"type": "null"
}
]
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 9
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"body_variant": {
"enum": [
"regular",
"extended",
"original"
],
"type": "string"
},
"do_not_share": {
"type": "boolean"
},
"document_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"draftpost_id": {
"anyOf": [
{
"type": "string",
"format": "uuid"
},
{
"type": "null"
}
]
},
"first_comment": {
"anyOf": [
{
"type": "string",
"maxLength": 1250
},
{
"type": "null"
}
]
},
"document_title": {
"anyOf": [
{
"type": "string",
"maxLength": 100,
"minLength": 1
},
{
"type": "null"
}
]
},
"first_comment_image_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"annotations": {
"title": "Publish Post",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description94/100
Schedule a post for future delivery to LinkedIn. Pass `draftpost_id` to hydrate text/image_urls/first_comment/video/document from the draft (defaults to the regular body — override with `body_variant`). Explicit fields take precedence. `text` is optional when media is attached: a media-only call schedules with no caption. The scheduled-posts cron picks it up at scheduled_time. `first_comment_image_url` attaches an image (or GIF) to the first comment. To schedule a LinkedIn poll, pass `poll` (question, options, duration) with the post text and no media; a draft that already carries a poll schedules it automatically. Returns a `url` deep link to the planner, where the user reviews and edits scheduled posts.
The description explains the capability, when to use it, and what to expect.
Inputs6/100
1 of 16 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
text
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
author_urn
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment_image_url
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
body_variant
Not provided
Missing description. Explain what this value means and how to supply it.
do_not_share
Not provided
Missing description. Explain what this value means and how to supply it.
document_url
Not provided
Missing description. Explain what this value means and how to supply it.
document_title
Not provided
Missing description. Explain what this value means and how to supply it.
video_url
Not provided
Missing description. Explain what this value means and how to supply it.
poll › question
Not provided
Missing description. Explain what this value means and how to supply it.
poll › options
Not provided
Missing description. Explain what this value means and how to supply it.
poll › duration
Not provided
Missing description. Explain what this value means and how to supply it.
scheduled_time
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "post_schedule",
"description": "Schedule a post for future delivery to LinkedIn. Pass `draftpost_id` to hydrate text/image_urls/first_comment/video/document from the draft (defaults to the regular body — override with `body_variant`). Explicit fields take precedence. `text` is optional when media is attached: a media-only call schedules with no caption. The scheduled-posts cron picks it up at scheduled_time. `first_comment_image_url` attaches an image (or GIF) to the first comment. To schedule a LinkedIn poll, pass `poll` (question, options, duration) with the post text and no media; a draft that already carries a poll schedules it automatically. Returns a `url` deep link to the planner, where the user reviews and edits scheduled posts.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"scheduled_time"
],
"properties": {
"poll": {
"type": "object",
"required": [
"question",
"options",
"duration"
],
"properties": {
"options": {
"type": "array",
"items": {
"type": "string",
"maxLength": 30,
"minLength": 1
},
"maxItems": 4,
"minItems": 2
},
"duration": {
"enum": [
"ONE_DAY",
"THREE_DAYS",
"SEVEN_DAYS",
"FOURTEEN_DAYS"
],
"type": "string"
},
"question": {
"type": "string",
"maxLength": 140,
"minLength": 1
}
},
"description": "A LinkedIn poll to publish with the post. Send it ONLY when the user asked for a poll. It needs a question and answer options: at least two options, at most four, each a short phrase. `duration` is how long voting stays open. A poll cannot be combined with images, a video or a document, and it cannot be edited after it is published. The post text still goes out above the poll.",
"additionalProperties": false
},
"text": {
"type": "string",
"minLength": 1,
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"video_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"author_urn": {
"anyOf": [
{
"type": "string",
"minLength": 1
},
{
"type": "null"
}
]
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 9
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"body_variant": {
"enum": [
"regular",
"extended",
"original"
],
"type": "string"
},
"do_not_share": {
"type": "boolean"
},
"document_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"draftpost_id": {
"anyOf": [
{
"type": "string",
"format": "uuid"
},
{
"type": "null"
}
]
},
"first_comment": {
"anyOf": [
{
"type": "string",
"maxLength": 1250
},
{
"type": "null"
}
]
},
"document_title": {
"anyOf": [
{
"type": "string",
"maxLength": 100,
"minLength": 1
},
{
"type": "null"
}
]
},
"scheduled_time": {
"type": "string",
"format": "date-time"
},
"first_comment_image_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"annotations": {
"title": "Schedule Post",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
Align the title with the capability actually described by this tool.
Description97/100
List posts for a profile from the planner queue, across LinkedIn, X and Telegram (each row has a `platform`). Defaults to pending rows only (`status=scheduled`, soonest first); pass `posted`, `failed`, or `all` for history (most recent first). Use this to find a `scheduled_post_id` before calling post_update (LinkedIn rows only) or post_delete (any platform). Filter with `draftpost_id` when the user has a draft open and you need its queued slot. `text_preview` is TRUNCATED for display only. Never pass it back as post_update `text`, or you will overwrite the post with a clipped copy. Pass `full_text: true` to also get each post's complete `text`, e.g. to review or edit the copy of queued posts.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 4 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
status
Not provided
Missing description. Explain what this value means and how to supply it.
full_text
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations25/100
Declare destructiveHint, idempotentHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "post_list",
"description": "List posts for a profile from the planner queue, across LinkedIn, X and Telegram (each row has a `platform`). Defaults to pending rows only (`status=scheduled`, soonest first); pass `posted`, `failed`, or `all` for history (most recent first). Use this to find a `scheduled_post_id` before calling post_update (LinkedIn rows only) or post_delete (any platform). Filter with `draftpost_id` when the user has a draft open and you need its queued slot. `text_preview` is TRUNCATED for display only. Never pass it back as post_update `text`, or you will overwrite the post with a clipped copy. Pass `full_text: true` to also get each post's complete `text`, e.g. to review or edit the copy of queued posts.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"status": {
"enum": [
"scheduled",
"posted",
"failed",
"all"
],
"type": "string"
},
"full_text": {
"type": "boolean"
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"draftpost_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "List Scheduled Posts",
"readOnlyHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description97/100
Update an existing scheduled LinkedIn post without cancelling and re-booking it. Pass `scheduled_post_id` from post_list (or from a prior post_schedule result). Provide only the fields that change: `text`, `image_urls`, `scheduled_time`, `first_comment`, video/document fields, or `poll` (a poll replaces any media on the queued post). Pass `draftpost_id` to pull the draft's current media onto the queued post (use this after draft_attachImages); it does NOT touch the queued text, which the user may have edited in the planner. To change copy, pass `text` explicitly. Use this when the user asks to add a photo, edit copy, or move the time on a post that is already in the planner. Does not work on posts that have already gone live. Returns the same planner deep link as post_schedule.
The description explains the capability, when to use it, and what to expect.
Inputs6/100
1 of 16 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
text
To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment.
Clear enough to supply this argument.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
author_urn
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment
Not provided
Missing description. Explain what this value means and how to supply it.
first_comment_image_url
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
do_not_share
Not provided
Missing description. Explain what this value means and how to supply it.
document_url
Not provided
Missing description. Explain what this value means and how to supply it.
document_title
Not provided
Missing description. Explain what this value means and how to supply it.
video_url
Not provided
Missing description. Explain what this value means and how to supply it.
poll › question
Not provided
Missing description. Explain what this value means and how to supply it.
poll › options
Not provided
Missing description. Explain what this value means and how to supply it.
poll › duration
Not provided
Missing description. Explain what this value means and how to supply it.
scheduled_post_id
Not provided
Missing description. Explain what this value means and how to supply it.
scheduled_time
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "post_update",
"description": "Update an existing scheduled LinkedIn post without cancelling and re-booking it. Pass `scheduled_post_id` from post_list (or from a prior post_schedule result). Provide only the fields that change: `text`, `image_urls`, `scheduled_time`, `first_comment`, video/document fields, or `poll` (a poll replaces any media on the queued post). Pass `draftpost_id` to pull the draft's current media onto the queued post (use this after draft_attachImages); it does NOT touch the queued text, which the user may have edited in the planner. To change copy, pass `text` explicitly. Use this when the user asks to add a photo, edit copy, or move the time on a post that is already in the planner. Does not work on posts that have already gone live. Returns the same planner deep link as post_schedule.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"scheduled_post_id"
],
"properties": {
"poll": {
"type": "object",
"required": [
"question",
"options",
"duration"
],
"properties": {
"options": {
"type": "array",
"items": {
"type": "string",
"maxLength": 30,
"minLength": 1
},
"maxItems": 4,
"minItems": 2
},
"duration": {
"enum": [
"ONE_DAY",
"THREE_DAYS",
"SEVEN_DAYS",
"FOURTEEN_DAYS"
],
"type": "string"
},
"question": {
"type": "string",
"maxLength": 140,
"minLength": 1
}
},
"description": "A LinkedIn poll to publish with the post. Send it ONLY when the user asked for a poll. It needs a question and answer options: at least two options, at most four, each a short phrase. `duration` is how long voting stays open. A poll cannot be combined with images, a video or a document, and it cannot be edited after it is published. The post text still goes out above the poll.",
"additionalProperties": false
},
"text": {
"type": "string",
"minLength": 1,
"description": "To tag a person or company, paste the `mention` string from linkedin_searchMentions inline where the name should appear (`@[Name](urn:li:person:X)` or `@[Name](urn:li:organization:N)`). A mention string may end with a quoted profile URL inside the parentheses; keep it exactly as given. Never invent a URN. Tags do not work in first_comment."
},
"video_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"author_urn": {
"anyOf": [
{
"type": "string",
"minLength": 1
},
{
"type": "null"
}
]
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 9
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"do_not_share": {
"type": "boolean"
},
"document_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
},
"draftpost_id": {
"anyOf": [
{
"type": "string",
"format": "uuid"
},
{
"type": "null"
}
]
},
"first_comment": {
"anyOf": [
{
"type": "string",
"maxLength": 1250
},
{
"type": "null"
}
]
},
"document_title": {
"anyOf": [
{
"type": "string",
"maxLength": 100,
"minLength": 1
},
{
"type": "null"
}
]
},
"scheduled_time": {
"type": "string",
"format": "date-time"
},
"scheduled_post_id": {
"type": "string",
"format": "uuid"
},
"first_comment_image_url": {
"anyOf": [
{
"type": "string",
"format": "uri"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"annotations": {
"title": "Update Scheduled Post",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description92/100
Cancel a scheduled post that hasn't been published yet — removes a queued post from the planner. This does NOT and cannot delete or remove a post that is already live on LinkedIn. Throws 404 if the post is already posted or does not exist.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
scheduled_post_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "post_delete",
"description": "Cancel a scheduled post that hasn't been published yet — removes a queued post from the planner. This does NOT and cannot delete or remove a post that is already live on LinkedIn. Throws 404 if the post is already posted or does not exist.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"scheduled_post_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
},
"scheduled_post_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Delete Scheduled Post",
"readOnlyHint": false,
"idempotentHint": true,
"destructiveHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description89/100
ONLY for X (Twitter), and ONLY when the user explicitly asked about X or Twitter. Connecting LinkedIn is a separate flow and this tool has no effect on it. Get a one-tap URL the user opens to connect (or reconnect) their X account to a 2pr profile. There is no X settings screen yet, so this URL is the only way to authorize. Give the user the link and ask them to open it; the page tells them when it worked. The link is single-use and valid for 7 days, so it still works if they come back to it later. Also reports whether X is already connected.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "x_connect",
"description": "ONLY for X (Twitter), and ONLY when the user explicitly asked about X or Twitter. Connecting LinkedIn is a separate flow and this tool has no effect on it. Get a one-tap URL the user opens to connect (or reconnect) their X account to a 2pr profile. There is no X settings screen yet, so this URL is the only way to authorize. Give the user the link and ask them to open it; the page tells them when it worked. The link is single-use and valid for 7 days, so it still works if they come back to it later. Also reports whether X is already connected.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Connect X Account",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description82/100
ONLY for X (Twitter), and ONLY when the user explicitly asked for X or Twitter. LinkedIn is the default destination for every post: if the user did not name X, use `post_publish` instead. A draft whose format is `x` was made for X, so publishing that draft here counts as the user naming it. Publishing to the wrong network cannot be undone. Publish to X immediately. Pass `text` for a single post or `thread` as an ordered array of segments for a thread. Links count as 23 characters regardless of their real length. The per-post limit depends on the connected account: 280 characters on a free account, up to 25000 on X Premium. Do NOT pre-split long content into a thread assuming 280: send it as one `text`, and only split if the tool rejects it for length. A thread of N segments costs N posts where one long post costs 1. Up to 4 images via `image_urls`, or the draft's one video (never both), attached to the first segment of a thread. Pass `draftpost_id` to hydrate the body from a draft; a draft that is too long is an error, never a silent trim. X has no document post type, no first comment and no company pages, so PDFs and carousels are rejected. Requires X to be connected: call `x_connect` first if it isn't. Set do_not_share:true to record the post without actually calling X. Scheduled X posts show in the 2pr planner next to LinkedIn; use `x_schedule` and `post_delete` to manage them.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 7 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
text
Not provided
Missing description. Explain what this value means and how to supply it.
thread
Not provided
Missing description. Explain what this value means and how to supply it.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
body_variant
Not provided
Missing description. Explain what this value means and how to supply it.
do_not_share
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "x_publish",
"description": "ONLY for X (Twitter), and ONLY when the user explicitly asked for X or Twitter. LinkedIn is the default destination for every post: if the user did not name X, use `post_publish` instead. A draft whose format is `x` was made for X, so publishing that draft here counts as the user naming it. Publishing to the wrong network cannot be undone. Publish to X immediately. Pass `text` for a single post or `thread` as an ordered array of segments for a thread. Links count as 23 characters regardless of their real length. The per-post limit depends on the connected account: 280 characters on a free account, up to 25000 on X Premium. Do NOT pre-split long content into a thread assuming 280: send it as one `text`, and only split if the tool rejects it for length. A thread of N segments costs N posts where one long post costs 1. Up to 4 images via `image_urls`, or the draft's one video (never both), attached to the first segment of a thread. Pass `draftpost_id` to hydrate the body from a draft; a draft that is too long is an error, never a silent trim. X has no document post type, no first comment and no company pages, so PDFs and carousels are rejected. Requires X to be connected: call `x_connect` first if it isn't. Set do_not_share:true to record the post without actually calling X. Scheduled X posts show in the 2pr planner next to LinkedIn; use `x_schedule` and `post_delete` to manage them.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"text": {
"type": "string",
"minLength": 1
},
"thread": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
},
"maxItems": 25,
"minItems": 1
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 4
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"body_variant": {
"enum": [
"regular",
"extended",
"original"
],
"type": "string"
},
"do_not_share": {
"type": "boolean"
},
"draftpost_id": {
"anyOf": [
{
"type": "string",
"format": "uuid"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"annotations": {
"title": "Publish to X",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description97/100
ONLY for X (Twitter), and ONLY when the user explicitly asked for X or Twitter. LinkedIn is the default destination for every post: if the user did not name X, use `post_schedule` instead. A draft whose format is `x` was made for X, so publishing that draft here counts as the user naming it. Schedule a post or thread for future delivery to X. Same content rules as `x_publish`: `text` (280 characters on a free account, up to 25000 on X Premium, so do not pre-split) or `thread` as an array of segments, up to 4 images or the draft's one video, no documents. The every-minute publish cron delivers it at `scheduled_time`. Scheduled X posts show in the 2pr planner, where the user can reschedule or cancel them; the returned `id` is what you pass to `post_delete` to cancel it before it goes out.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 8 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
text
Not provided
Missing description. Explain what this value means and how to supply it.
thread
Not provided
Missing description. Explain what this value means and how to supply it.
image_urls
Not provided
Missing description. Explain what this value means and how to supply it.
draftpost_id
Not provided
Missing description. Explain what this value means and how to supply it.
body_variant
Not provided
Missing description. Explain what this value means and how to supply it.
do_not_share
Not provided
Missing description. Explain what this value means and how to supply it.
scheduled_time
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "x_schedule",
"description": "ONLY for X (Twitter), and ONLY when the user explicitly asked for X or Twitter. LinkedIn is the default destination for every post: if the user did not name X, use `post_schedule` instead. A draft whose format is `x` was made for X, so publishing that draft here counts as the user naming it. Schedule a post or thread for future delivery to X. Same content rules as `x_publish`: `text` (280 characters on a free account, up to 25000 on X Premium, so do not pre-split) or `thread` as an array of segments, up to 4 images or the draft's one video, no documents. The every-minute publish cron delivers it at `scheduled_time`. Scheduled X posts show in the 2pr planner, where the user can reschedule or cancel them; the returned `id` is what you pass to `post_delete` to cancel it before it goes out.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"scheduled_time"
],
"properties": {
"text": {
"type": "string",
"minLength": 1
},
"thread": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
},
"maxItems": 25,
"minItems": 1
},
"image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 4
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"body_variant": {
"enum": [
"regular",
"extended",
"original"
],
"type": "string"
},
"do_not_share": {
"type": "boolean"
},
"draftpost_id": {
"anyOf": [
{
"type": "string",
"format": "uuid"
},
{
"type": "null"
}
]
},
"scheduled_time": {
"type": "string",
"format": "date-time"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Schedule to X",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
ONLY for X (Twitter), and ONLY when the user asked about X or Twitter. The analytics_* tools are LinkedIn-only; this is the tool for X. Fetch live stats for the connected X account: followers, following, total post count, and the most recent original posts (replies and reposts excluded) with impressions, likes, reposts, replies, quotes, bookmarks and a link for each. Default 20 posts, 5 to 100 via `limit`. Read live from X on every call, so there is no history over time and no follower trend: say so if the user asks for growth over a period. Requires X to be connected: call `x_connect` first if it isn't.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
limit
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare destructiveHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "x_getStats",
"description": "ONLY for X (Twitter), and ONLY when the user asked about X or Twitter. The analytics_* tools are LinkedIn-only; this is the tool for X. Fetch live stats for the connected X account: followers, following, total post count, and the most recent original posts (replies and reposts excluded) with impressions, likes, reposts, replies, quotes, bookmarks and a link for each. Default 20 posts, 5 to 100 via `limit`. Read live from X on every call, so there is no history over time and no follower trend: say so if the user asks for growth over a period. Requires X to be connected: call `x_connect` first if it isn't.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"limit": {
"type": "integer",
"maximum": 100,
"minimum": 5
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get X Stats",
"readOnlyHint": true,
"openWorldHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description96/100
Upload an image, video, or PDF document (base64-encoded) to 2pr storage and return its public URL and media `kind`. Supported content types: image/png, image/jpeg, image/webp, image/gif (<= 10 MB); video/mp4, video/quicktime (<= 100 MB); application/pdf (<= 100 MB). The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title). For video and PDF, prefer the two-step media_prepareUpload/media_commitUpload flow — files this large rarely fit inside a single tool call as base64.
The description explains the capability, when to use it, and what to expect.
Inputs30/100
2 of 5 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
filename
Not provided
Missing description. Explain what this value means and how to supply it.
content_type
Not provided
Missing description. Explain what this value means and how to supply it.
data_base64
Raw base64 (no `data:` prefix). Decoded byte size must be within the per-type limit.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
alt_text
Alt text for images; used as the document title for PDFs.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
Annotations50/100
Declare idempotentHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "media_upload",
"description": "Upload an image, video, or PDF document (base64-encoded) to 2pr storage and return its public URL and media `kind`. Supported content types: image/png, image/jpeg, image/webp, image/gif (<= 10 MB); video/mp4, video/quicktime (<= 100 MB); application/pdf (<= 100 MB). The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title). For video and PDF, prefer the two-step media_prepareUpload/media_commitUpload flow — files this large rarely fit inside a single tool call as base64.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"filename",
"content_type",
"data_base64"
],
"properties": {
"alt_text": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
],
"description": "Alt text for images; used as the document title for PDFs."
},
"filename": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"data_base64": {
"type": "string",
"minLength": 1,
"description": "Raw base64 (no `data:` prefix). Decoded byte size must be within the per-type limit."
},
"content_type": {
"enum": [
"image/png",
"image/jpeg",
"image/jpg",
"image/webp",
"image/gif",
"video/mp4",
"video/quicktime",
"application/pdf"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Upload Media",
"readOnlyHint": false,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Fetch a media file from a public URL and re-upload it to 2pr storage. Returns the rehosted public URL and media `kind`. Use this when the user has the file at any public link (imgbb, Dropbox share, an existing CDN URL) and the agent cannot read the raw bytes itself. Allowed content types: image/png|jpeg|webp|gif (<= 10 MB); video/mp4, video/quicktime (<= 100 MB); application/pdf (<= 100 MB). The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title).
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 4 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
url
Not provided
Missing description. Explain what this value means and how to supply it.
filename
Not provided
Missing description. Explain what this value means and how to supply it.
alt_text
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
{
"title": "Upload Media from URL",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
Output schema0/100
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "media_uploadFromUrl",
"description": "Fetch a media file from a public URL and re-upload it to 2pr storage. Returns the rehosted public URL and media `kind`. Use this when the user has the file at any public link (imgbb, Dropbox share, an existing CDN URL) and the agent cannot read the raw bytes itself. Allowed content types: image/png|jpeg|webp|gif (<= 10 MB); video/mp4, video/quicktime (<= 100 MB); application/pdf (<= 100 MB). The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title).",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"url"
],
"properties": {
"url": {
"type": "string",
"format": "uri"
},
"alt_text": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
]
},
"filename": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Upload Media from URL",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description93/100
Step 1 of the two-step upload flow. Returns a signed PUT URL the agent can upload bytes to out-of-band (no base64 in context). After the PUT succeeds, call media_commitUpload with `storage_path` to finalise and get the public URL. Use this for any local file — it scales to the full per-type ceiling (10 MB images, 100 MB video/PDF), unlike media_upload which is bottlenecked by tool-call context limits. Supported content types: image/png|jpeg|webp|gif, video/mp4, video/quicktime, application/pdf.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
content_type
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "media_prepareUpload",
"description": "Step 1 of the two-step upload flow. Returns a signed PUT URL the agent can upload bytes to out-of-band (no base64 in context). After the PUT succeeds, call media_commitUpload with `storage_path` to finalise and get the public URL. Use this for any local file — it scales to the full per-type ceiling (10 MB images, 100 MB video/PDF), unlike media_upload which is bottlenecked by tool-call context limits. Supported content types: image/png|jpeg|webp|gif, video/mp4, video/quicktime, application/pdf.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"content_type"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
},
"content_type": {
"enum": [
"image/png",
"image/jpeg",
"image/jpg",
"image/webp",
"image/gif",
"video/mp4",
"video/quicktime",
"application/pdf"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Prepare Media Upload",
"readOnlyHint": false,
"idempotentHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description97/100
Step 2 of the two-step upload flow. Pass the `storage_path` returned by media_prepareUpload after your PUT upload completed. Server verifies the object exists, size is within the per-type limit, and content type is allowed, then returns the public URL and media `kind`. The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title).
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 3 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
storage_path
Not provided
Missing description. Explain what this value means and how to supply it.
alt_text
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "media_commitUpload",
"description": "Step 2 of the two-step upload flow. Pass the `storage_path` returned by media_prepareUpload after your PUT upload completed. Server verifies the object exists, size is within the per-type limit, and content type is allowed, then returns the public URL and media `kind`. The returned `kind` tells you where the URL goes: attach to a draft with draft_attachImages (`image`) or draft_attachMedia (`video`/`document`), or pass it straight to post_publish/post_schedule as `image_urls` (image), `video_url` (video), or `document_url` + `document_title` (document — the `alt_text` you supply is used as the title).",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"storage_path"
],
"properties": {
"alt_text": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
]
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"storage_path": {
"type": "string",
"maxLength": 500,
"minLength": 1
}
},
"additionalProperties": false
},
"annotations": {
"title": "Commit Media Upload",
"readOnlyHint": false,
"idempotentHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description97/100
Generate an AI image from a text prompt (the same "create image" pipeline the draft editor uses) and return a hosted 2pr public URL. Pass that url to draft_attachImages, or in `image_urls` on post_publish/post_schedule. `prompt` should describe the image you want (the generator targets a 4:5 portrait aspect ratio suited to LinkedIn). Optional `reference_image_urls` (up to 3 public image urls, <= 10 MB each) edits, varies, or combines those images instead of generating from scratch — pass several when the user gives you several (two brand logos to put in one scene, a screenshot plus a style reference), and describe in `prompt` what to do with each. An image the user attached in chat is already hosted: use its URL here rather than describing it from memory. When the post is built on a source (draft_create `source_page_image_urls`, web_fetch `image_urls`) and the user wants a picture or versions of the picture, pass the source picture in `reference_image_urls` so each version is a variation of the original rather than a guess from the text. `backend` defaults to a fast, cheap model for fresh generations and to `google_banana_pro` (best edit quality) when reference images are set; an explicit `backend` is always honored, so pass `openai` if the user asks for the OpenAI/ChatGPT image model. Requires an active subscription; rate-limited.
The description explains the capability, when to use it, and what to expect.
Inputs36/100
3 of 7 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
prompt
Not provided
Missing description. Explain what this value means and how to supply it.
backend
Not provided
Missing description. Explain what this value means and how to supply it.
reference_image_urls
Public image URLs to edit, vary, or combine. Prefer this over reference_image_url.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
reference_image_url
Deprecated single-image form of reference_image_urls; still accepted.
Clear enough to supply this argument.
alt_text
Not provided
Missing description. Explain what this value means and how to supply it.
orientation
portrait (4:5, default) for feed posts; landscape (16:9) for a LinkedIn article cover.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "media_generateImage",
"description": "Generate an AI image from a text prompt (the same \"create image\" pipeline the draft editor uses) and return a hosted 2pr public URL. Pass that url to draft_attachImages, or in `image_urls` on post_publish/post_schedule. `prompt` should describe the image you want (the generator targets a 4:5 portrait aspect ratio suited to LinkedIn). Optional `reference_image_urls` (up to 3 public image urls, <= 10 MB each) edits, varies, or combines those images instead of generating from scratch — pass several when the user gives you several (two brand logos to put in one scene, a screenshot plus a style reference), and describe in `prompt` what to do with each. An image the user attached in chat is already hosted: use its URL here rather than describing it from memory. When the post is built on a source (draft_create `source_page_image_urls`, web_fetch `image_urls`) and the user wants a picture or versions of the picture, pass the source picture in `reference_image_urls` so each version is a variation of the original rather than a guess from the text. `backend` defaults to a fast, cheap model for fresh generations and to `google_banana_pro` (best edit quality) when reference images are set; an explicit `backend` is always honored, so pass `openai` if the user asks for the OpenAI/ChatGPT image model. Requires an active subscription; rate-limited.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"prompt"
],
"properties": {
"prompt": {
"type": "string",
"maxLength": 2000,
"minLength": 1
},
"backend": {
"enum": [
"google_banana_lite",
"google_banana_pro",
"openai"
],
"type": "string"
},
"alt_text": {
"anyOf": [
{
"type": "string",
"maxLength": 500
},
{
"type": "null"
}
]
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"orientation": {
"enum": [
"portrait",
"landscape"
],
"type": "string",
"description": "portrait (4:5, default) for feed posts; landscape (16:9) for a LinkedIn article cover."
},
"reference_image_url": {
"type": "string",
"format": "uri",
"description": "Deprecated single-image form of reference_image_urls; still accepted."
},
"reference_image_urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 3,
"minItems": 1,
"description": "Public image URLs to edit, vary, or combine. Prefer this over reference_image_url."
}
},
"additionalProperties": false
},
"annotations": {
"title": "Generate Image",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description92/100
Kick off background enrichment of the caller's active profile from a LinkedIn URL (pulls headline, summary, recent posts, etc.). Returns the current profile snapshot immediately — the enriched fields land asynchronously.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
linkedin_url
Not provided
Missing description. Explain what this value means and how to supply it.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Resolve a person or company to the LinkedIn URN needed to @-mention them in a post. Query with a full name, a company name, or a LinkedIn profile or company page URL. Each result carries a ready-to-use `mention` string: paste it verbatim into the post `text` where the tag should appear and LinkedIn renders a real tag on publish. Tags only work in the post body, not in `first_comment`. Coverage is partial: a person is found only if they follow a page the user administers, have engaged with the user, or their profile URL is passed directly; a company is found only by an exact page name or URL. Needs the user's LinkedIn account connected. An empty list means no match: if the query was a name, ask the user for the profile URL once; otherwise post without the tag. Never invent a URN, an unknown one publishes as literal text.
The description explains the capability, when to use it, and what to expect.
Inputs39/100
2 of 3 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
query
A full name, a company name, or a LinkedIn profile or company page URL.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
type
Limit the search to people or companies. Omit to search both.
Clear enough to supply this argument.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "linkedin_searchMentions",
"description": "Resolve a person or company to the LinkedIn URN needed to @-mention them in a post. Query with a full name, a company name, or a LinkedIn profile or company page URL. Each result carries a ready-to-use `mention` string: paste it verbatim into the post `text` where the tag should appear and LinkedIn renders a real tag on publish. Tags only work in the post body, not in `first_comment`. Coverage is partial: a person is found only if they follow a page the user administers, have engaged with the user, or their profile URL is passed directly; a company is found only by an exact page name or URL. Needs the user's LinkedIn account connected. An empty list means no match: if the query was a name, ask the user for the profile URL once; otherwise post without the tag. Never invent a URN, an unknown one publishes as literal text.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"query"
],
"properties": {
"type": {
"enum": [
"person",
"company"
],
"type": "string",
"description": "Limit the search to people or companies. Omit to search both."
},
"query": {
"type": "string",
"maxLength": 200,
"minLength": 1,
"description": "A full name, a company name, or a LinkedIn profile or company page URL."
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Search LinkedIn Mentions",
"readOnlyHint": false,
"openWorldHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description95/100
Fetch the user's LinkedIn engagement inbox: top-level comments on their recent posts (last 30 days, up to 50 posts), with parent post context and a `needs_reply` flag. Served from a stored copy that refreshes when the user opens the Engage page, not on every call; `freshness_note` in the result says how old the copy is, and a very recent comment may not be in it yet. Set only_needs_reply:true to filter out comments the user has already replied to.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
only_needs_reply
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engage_getInbox",
"description": "Fetch the user's LinkedIn engagement inbox: top-level comments on their recent posts (last 30 days, up to 50 posts), with parent post context and a `needs_reply` flag. Served from a stored copy that refreshes when the user opens the Engage page, not on every call; `freshness_note` in the result says how old the copy is, and a very recent comment may not be in it yet. Set only_needs_reply:true to filter out comments the user has already replied to.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
},
"only_needs_reply": {
"type": "boolean"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Engagement Inbox",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description81/100
Generate 5 AI reply variants for a LinkedIn comment, using the user's persona and custom engage instructions. Each reply is capped at 220 characters. Pass the comment text and (optionally) the parent post text and commenter identity for better grounding.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 5 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
comment_text
Not provided
Missing description. Explain what this value means and how to supply it.
post_text
Not provided
Missing description. Explain what this value means and how to supply it.
commenter_name
Not provided
Missing description. Explain what this value means and how to supply it.
commenter_headline
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, idempotentHint based on the tool’s actual behavior.
Align the title with the capability actually described by this tool.
Description97/100
Post a reply to a LinkedIn post or comment. Omit parent_comment_urn to post a top-level comment on the share; provide it (a urn:li:comment:(...) URN) to reply within a comment thread. Returns the new comment's URN.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 4 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
share_urn
Not provided
Missing description. Explain what this value means and how to supply it.
text
Not provided
Missing description. Explain what this value means and how to supply it.
parent_comment_urn
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations75/100
Declare idempotentHint based on the tool’s actual behavior.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description84/100
Set or clear the caller's reaction to a LinkedIn comment. Pass a reaction_type to react (LIKE, PRAISE, EMPATHY, INTEREST, APPRECIATION, ENTERTAINMENT); pass null to clear an existing reaction. Idempotent: setting the same reaction twice is a no-op.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 3 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
comment_urn
Not provided
Missing description. Explain what this value means and how to supply it.
reaction_type
Not provided
Missing description. Explain what this value means and how to supply it.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description98/100
Refresh the user's LinkedIn posts and their comments from LinkedIn (what the engage inbox reads). Slow — can take 2-3 minutes. Use when the inbox looks stale or after the user posts off-platform. Also marks the profile as synced so the on-load auto-refresh won't re-fire. Returns the number of posts upserted.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engage_refreshPosts",
"description": "Refresh the user's LinkedIn posts and their comments from LinkedIn (what the engage inbox reads). Slow — can take 2-3 minutes. Use when the inbox looks stale or after the user posts off-platform. Also marks the profile as synced so the on-load auto-refresh won't re-fire. Returns the number of posts upserted.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Refresh Engagement Posts",
"readOnlyHint": false,
"openWorldHint": true,
"idempotentHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description89/100
List the "Engage with" prospect list on the Engage page (Outbound comments): the people whose posts the user tracks to comment on. Separate from the "My Creators" list (creator_list), which feeds the viral library. Returns the prospects plus how many slots the list has and how many are still free.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "outbound_listProspects",
"description": "List the \"Engage with\" prospect list on the Engage page (Outbound comments): the people whose posts the user tracks to comment on. Separate from the \"My Creators\" list (creator_list), which feeds the viral library. Returns the prospects plus how many slots the list has and how many are still free.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "List Prospects",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description88/100
Add a person to the "Engage with" prospect list (Outbound comments) by LinkedIn member profile URL or bare vanity handle. A plain name only works when it matches exactly one creator we already index, so prefer the profile URL. Company pages are rejected. Adding someone already on the list returns the existing entry. The list has a fixed number of slots; outbound_listProspects reports how many are left.
Describe the returned information so the agent knows whether it will answer the request.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
linkedin_url
Not provided
Missing description. Explain what this value means and how to supply it.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "outbound_addProspect",
"description": "Add a person to the \"Engage with\" prospect list (Outbound comments) by LinkedIn member profile URL or bare vanity handle. A plain name only works when it matches exactly one creator we already index, so prefer the profile URL. Company pages are rejected. Adding someone already on the list returns the existing entry. The list has a fixed number of slots; outbound_listProspects reports how many are left.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id",
"linkedin_url"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
},
"linkedin_url": {
"type": "string",
"maxLength": 500,
"minLength": 1
}
},
"additionalProperties": false
},
"annotations": {
"title": "Add Prospect",
"readOnlyHint": false,
"openWorldHint": true,
"idempotentHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description93/100
Remove a prospect from the "Engage with" list (Outbound comments) by the id outbound_listProspects returns. Frees a slot; their posts stop appearing in the outbound feed.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
prospect_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare idempotentHint, openWorldHint based on the tool’s actual behavior.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description87/100
Recent LinkedIn posts by the people on the "Engage with" prospect list (Outbound comments), newest first: the last few posts per prospect from the past 5 days, with engagement counts and whether the user already commented. Pass prospect_id to get one person's latest posts. Served from a stored copy that refreshes when the user opens the Outbound tab; it never contacts LinkedIn. `freshness_note` says how old the copy is; when `stale_prospects` is above zero, call outbound_refreshFeed and read again before concluding someone has not posted.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
prospect_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "outbound_getFeed",
"description": "Recent LinkedIn posts by the people on the \"Engage with\" prospect list (Outbound comments), newest first: the last few posts per prospect from the past 5 days, with engagement counts and whether the user already commented. Pass prospect_id to get one person's latest posts. Served from a stored copy that refreshes when the user opens the Outbound tab; it never contacts LinkedIn. `freshness_note` says how old the copy is; when `stale_prospects` is above zero, call outbound_refreshFeed and read again before concluding someone has not posted.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
},
"prospect_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Prospect Posts",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description91/100
Poll LinkedIn for new posts from the "Engage with" prospects whose stored copy is older than the refresh window, then read outbound_getFeed again for the result. Can take up to a minute for a full list. Each prospect polled spends provider budget, so there is a daily cap; `refresh_limited` is true when some prospects could not be refreshed today and the feed still holds their older posts. Set force:true to re-poll everyone regardless of age; it still counts against the cap.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 2 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
force
Not provided
Missing description. Explain what this value means and how to supply it.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "outbound_refreshFeed",
"description": "Poll LinkedIn for new posts from the \"Engage with\" prospects whose stored copy is older than the refresh window, then read outbound_getFeed again for the result. Can take up to a minute for a full list. Each prospect polled spends provider budget, so there is a daily cap; `refresh_limited` is true when some prospects could not be refreshed today and the feed still holds their older posts. Set force:true to re-poll everyone regardless of age; it still counts against the cap.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"force": {
"type": "boolean"
},
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Refresh Prospect Posts",
"readOnlyHint": false,
"openWorldHint": true,
"idempotentHint": true,
"destructiveHint": false
}
}
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engagers_listWatches",
"description": "List the LinkedIn and X profiles this account tracks for the engager feed (the people and company pages whose post engagers are collected every day).",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {},
"additionalProperties": false
},
"annotations": {
"title": "List Tracked Profiles",
"readOnlyHint": true,
"idempotentHint": true
}
}
Name the specific action and subject instead of using a generic label such as ‘manage’ or ‘tool’.
Description86/100
Start tracking a profile for the engager feed: a LinkedIn person (https://www.linkedin.com/in/name/), a LinkedIn company page (URL plus its numeric company id), or an X account (URL or @handle). The first daily collection looks back 30 days; after that each day adds new engagers. Adding a profile already tracked returns it unchanged.
The description explains the capability, when to use it, and what to expect.
Inputs56/100
3 of 4 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
platform
Not provided
Missing description. Explain what this value means and how to supply it.
profile
LinkedIn profile or company URL, or X profile URL or @handle.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
display_name
Name to show in the feed, e.g. "Rory O'Driscoll".
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
company_id
LinkedIn company pages only: the numeric company id.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
Annotations75/100
Declare openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engagers_addWatch",
"description": "Start tracking a profile for the engager feed: a LinkedIn person (https://www.linkedin.com/in/name/), a LinkedIn company page (URL plus its numeric company id), or an X account (URL or @handle). The first daily collection looks back 30 days; after that each day adds new engagers. Adding a profile already tracked returns it unchanged.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"platform",
"profile"
],
"properties": {
"profile": {
"type": "string",
"maxLength": 300,
"minLength": 1,
"description": "LinkedIn profile or company URL, or X profile URL or @handle."
},
"platform": {
"enum": [
"linkedin",
"x"
],
"type": "string"
},
"company_id": {
"type": "string",
"maxLength": 30,
"description": "LinkedIn company pages only: the numeric company id."
},
"display_name": {
"type": "string",
"maxLength": 200,
"description": "Name to show in the feed, e.g. \"Rory O'Driscoll\"."
}
},
"additionalProperties": false
},
"annotations": {
"title": "Track Profile",
"readOnlyHint": false,
"idempotentHint": true,
"destructiveHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description94/100
Read the engager feed: one row per person per engagement (reaction, comment, repost on LinkedIn; reply, repost, quote on X) on the tracked profiles' posts, plus new followers of tracked X accounts (engagement "follow"), newest first. LinkedIn rows include the person's clean profile URL, job title, company and location once looked up. Page with next_cursor. For a spreadsheet or CRM import, use engagers_exportCsv instead.
The description explains the capability, when to use it, and what to expect.
Inputs64/100
5 of 7 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
since
Only rows first collected on or after this date or ISO time, e.g. 2026-11-10. Use it to pull only what is new since the last pull.
Clear enough to supply this argument.
until
Only rows first collected before this date or ISO time.
Clarify units, where to obtain this value, or conventions not already expressed in the schema.
platform
Not provided
Missing description. Explain what this value means and how to supply it.
watch_id
Only engagements on one tracked profile (id from engagers_listWatches).
Clear enough to supply this argument.
engagement_type
Not provided
Missing description. Explain what this value means and how to supply it.
limit
Rows per page. Default 100.
Clear enough to supply this argument.
cursor
next_cursor from the previous page.
Clear enough to supply this argument.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engagers_list",
"description": "Read the engager feed: one row per person per engagement (reaction, comment, repost on LinkedIn; reply, repost, quote on X) on the tracked profiles' posts, plus new followers of tracked X accounts (engagement \"follow\"), newest first. LinkedIn rows include the person's clean profile URL, job title, company and location once looked up. Page with next_cursor. For a spreadsheet or CRM import, use engagers_exportCsv instead.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {
"limit": {
"type": "integer",
"maximum": 500,
"minimum": 1,
"description": "Rows per page. Default 100."
},
"since": {
"type": "string",
"description": "Only rows first collected on or after this date or ISO time, e.g. 2026-11-10. Use it to pull only what is new since the last pull."
},
"until": {
"type": "string",
"description": "Only rows first collected before this date or ISO time."
},
"cursor": {
"type": "string",
"format": "uuid",
"description": "next_cursor from the previous page."
},
"platform": {
"enum": [
"linkedin",
"x"
],
"type": "string"
},
"watch_id": {
"type": "string",
"format": "uuid",
"description": "Only engagements on one tracked profile (id from engagers_listWatches)."
},
"engagement_type": {
"enum": [
"reaction",
"comment",
"repost",
"reply",
"quote",
"like",
"follow"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Read Engager Feed",
"readOnlyHint": true,
"idempotentHint": true
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description92/100
Get a CSV download link for the engager feed, with the same filters as engagers_list. The link works without logging in and expires after 24 hours. Give the user the url.
The description explains the capability, when to use it, and what to expect.
Inputs60/100
3 of 5 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
since
Only rows first collected on or after this date or ISO time, e.g. 2026-11-10. Use it to pull only what is new since the last pull.
Clear enough to supply this argument.
until
Only rows first collected before this date or ISO time.
Clear enough to supply this argument.
platform
Not provided
Missing description. Explain what this value means and how to supply it.
watch_id
Only engagements on one tracked profile (id from engagers_listWatches).
Clear enough to supply this argument.
engagement_type
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "engagers_exportCsv",
"description": "Get a CSV download link for the engager feed, with the same filters as engagers_list. The link works without logging in and expires after 24 hours. Give the user the url.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {
"since": {
"type": "string",
"description": "Only rows first collected on or after this date or ISO time, e.g. 2026-11-10. Use it to pull only what is new since the last pull."
},
"until": {
"type": "string",
"description": "Only rows first collected before this date or ISO time."
},
"platform": {
"enum": [
"linkedin",
"x"
],
"type": "string"
},
"watch_id": {
"type": "string",
"format": "uuid",
"description": "Only engagements on one tracked profile (id from engagers_listWatches)."
},
"engagement_type": {
"enum": [
"reaction",
"comment",
"repost",
"reply",
"quote",
"like",
"follow"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Export Engager Feed CSV",
"readOnlyHint": true,
"idempotentHint": false
}
}
The title clearly communicates the tool’s action and subject. Keep it concise.
Description93/100
Fetch the profile's LinkedIn analytics dashboard — current follower count, daily follower changes, daily impressions/reactions/comments, members reached total, and daily post counts. Works for personal and company profiles. Default window is the last 730 days (personal) or 365 days (company, LinkedIn cap). Requires an active Pro/Team subscription. LinkedIn only: for X (Twitter) stats use `x_getStats`.
The description explains the capability, when to use it, and what to expect.
Inputs50/100
2 of 3 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
start_date
ISO 8601 timestamp. Defaults to ~2 years ago.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
end_date
ISO 8601 timestamp. Defaults to now.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
The title clearly communicates the tool’s action and subject. Keep it concise.
Description96/100
Per-post analytics for the profile's LinkedIn posts: likes, comments, post type, URL, posted timestamp, and time-series metrics snapshot (impressions, members reached, reactions, saves, clicks). If the stored metrics are older than 6 hours, the call first refreshes them from LinkedIn, which can take 10-40 seconds. Supports pagination and sorting; sort_by impressions ranks top posts by lifetime impressions, with posts that have no impression count last. Each post carries funnel_stage: tofu (visibility: opinions, takes, personal updates), mofu (education: how-tos, frameworks, case studies, lessons), bofu (explicit call to action), or null if not classified yet. Pass funnel_stage to only return posts of that stage, e.g. mofu for the user's educational posts. Requires an active Pro/Team subscription. LinkedIn only: for X (Twitter) posts use `x_getStats`.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 8 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
start_date
Not provided
Missing description. Explain what this value means and how to supply it.
end_date
Not provided
Missing description. Explain what this value means and how to supply it.
offset
Not provided
Missing description. Explain what this value means and how to supply it.
limit
Not provided
Missing description. Explain what this value means and how to supply it.
sort_by
Not provided
Missing description. Explain what this value means and how to supply it.
sort_direction
Not provided
Missing description. Explain what this value means and how to supply it.
funnel_stage
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
{
"title": "Get Post Analytics",
"readOnlyHint": true,
"idempotentHint": true
}
Output schema0/100
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "analytics_getPostStats",
"description": "Per-post analytics for the profile's LinkedIn posts: likes, comments, post type, URL, posted timestamp, and time-series metrics snapshot (impressions, members reached, reactions, saves, clicks). If the stored metrics are older than 6 hours, the call first refreshes them from LinkedIn, which can take 10-40 seconds. Supports pagination and sorting; sort_by impressions ranks top posts by lifetime impressions, with posts that have no impression count last. Each post carries funnel_stage: tofu (visibility: opinions, takes, personal updates), mofu (education: how-tos, frameworks, case studies, lessons), bofu (explicit call to action), or null if not classified yet. Pass funnel_stage to only return posts of that stage, e.g. mofu for the user's educational posts. Requires an active Pro/Team subscription. LinkedIn only: for X (Twitter) posts use `x_getStats`.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"limit": {
"type": "integer",
"maximum": 100,
"minimum": 1
},
"offset": {
"type": "integer",
"minimum": 0
},
"sort_by": {
"enum": [
"created_at_linkedin",
"likes",
"comments",
"impressions"
],
"type": "string"
},
"end_date": {
"type": "string",
"format": "date-time"
},
"profile_id": {
"type": "string",
"format": "uuid"
},
"start_date": {
"type": "string",
"format": "date-time"
},
"funnel_stage": {
"enum": [
"tofu",
"mofu",
"bofu"
],
"type": "string"
},
"sort_direction": {
"enum": [
"asc",
"desc"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Post Analytics",
"readOnlyHint": true,
"idempotentHint": true
}
}
Align the title with the capability actually described by this tool.
Description92/100
Posting-time history for the profile over the last 365 days — ISO timestamps of every LinkedIn post. Use to build best-time-to-post heatmaps. Requires an active Pro/Team subscription.
The description explains the capability, when to use it, and what to expect.
Inputs0/100
0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
profile_id
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations50/100
Declare destructiveHint, openWorldHint based on the tool’s actual behavior.
No output schema is declared. Add one for structured results so clients can interpret returned fields.
Not declared
Original tool metadata
{
"name": "analytics_getPostingTimes",
"description": "Posting-time history for the profile over the last 365 days — ISO timestamps of every LinkedIn post. Use to build best-time-to-post heatmaps. Requires an active Pro/Team subscription.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"profile_id"
],
"properties": {
"profile_id": {
"type": "string",
"format": "uuid"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Get Best Posting Times",
"readOnlyHint": true,
"idempotentHint": true
}
}
Loading score analytics…
Put your MCP to the test
A few tasks worth trying with this server, chosen from its reviewed tools. These have not been run.
01Search Viral PostsSearch the viral LinkedIn post catalog by keyword, with optional filters (language, creator URLs, engagement thresholds, date range).
02List ProfilesList all profiles the caller has access to.
03Draft Comment RepliesGenerate 5 AI reply variants for a LinkedIn comment, using the user's persona and custom engage instructions.
04Get DraftRead an existing draft by id.
05List Tracked ProfilesList the LinkedIn and X profiles this account tracks for the engager feed (the people and company pages whose post engagers are collected every day).
Want to run these tasks against your MCP server with us?
2pr.io setup, tools, quality score, publishing readiness, and alternatives.
How do I add 2pr.io to Claude, ChatGPT, or Cursor?
Add 2pr.io to Claude, ChatGPT, or Cursor by connecting its remote MCP URL in each client's MCP settings. In Claude, add a custom connector. In ChatGPT, add an MCP app. In Cursor, add an MCP server under Settings. The MCP URL is https://2pr.io/api/mcp.
Is 2pr.io MCP ready for ChatGPT and Claude?
As of November 15, 2026, 2pr.io scores 90/100 on Manufact's MCP publishing checklist. The scan does not yet meet ChatGPT publishing checks and meets Claude publishing checks. Passing these checks does not guarantee marketplace approval.
How clear is 2pr.io's MCP tool documentation?
As of November 15, 2026, 2pr.io scores 60/100 in Manufact's review of tool titles, descriptions, input and output schemas, and behavior annotations. The review covers 51 of 51 discovered tools. The quality score measures documentation clarity; runtime behavior and marketplace readiness require separate checks.
What are the best alternatives to 2pr.io?
As of November 16, 2026, alternatives to 2pr.io in marketing-seo on the Manufact MCP leaderboard include Notra, BulkPublish, Fuxux, HYPD AI for App Store Optimization, and HYPD AI for Google Search Console.
How can 2pr.io improve its MCP documentation?
creator_list is the lowest-scoring reviewed tool at 50/100. 0 of 1 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values. Open its review for feedback and a copyable fix prompt, then rerun the documentation review after updating your tool metadata.
Not your MCP?
Review any public or OAuth MCP server for protocol compatibility, tool documentation, and ChatGPT and Claude publishing readiness. No account required.