Vyrgent Skills & Plugin Hub

Searches X (Twitter) for a keyword/hashtag and extracts a batch of posts (author, text, stable
postldentffier/permalink) from the results feed - all in ONE tool call, no manual navigate/scroll/read_page loop needed. Use this FIRST when asked to find and reply to X posts about a topic. After getting results, compose a unique reply for each post you want to engage with, then call browser_agent's post_comment_safely (platform: "x") once per post using the postldentifier returned here - this tool does NOT post anything itself.

Publishes a brand-new top-level post on X (Twitter) -NOT a reply to an existing post. Give it the finished post text and it handles navigating to X, opening the composer, typing, and submitting in one hardcoded tool call. For replying to an existing post found via x_search, use browser_agent's post_comment_safely (platform: "x") instead - this tool does not reply to anything.

Uploads a focal video or photo file to TikTok and publishes it as a post with the given caption - navigate, file injection {no dialog), upload-processing wait, caption typing, and Post click alt in ONE hardcoded tool call. Use when the user sends a video/picture and asks to post it to TikTok: compose the caption yourself (they may want hashtags}, then call this with the file's local path and the finished caption. Requires the user to be logged into TikTok in the Vyrgent browser.

Creates a realistic lip-synced TALKING video from a photo of a person plus a spoken script, using the D-10 API - image upload, generation, polling, and mp4 download all in ONE hardcoded call. Returns the local mp4 path (hand it to tiktok_post/facebook_post etc. afterwards). Write the script yourself as natural SPOKEN language (short sentences, ~30-60 seconds when read aloud - roughly 80-150 words); do not include hashtags or stage directions in the script. Uses the user's configured avatar photo by default (didAvatarlmage), or pass imagePath to override. Requires didApiKey in config.

Automates Simplifile.com document submission. Logs in, selects county, enters package name, selects document type, adds document, uploads the file the user attached in chat, then submits. If the user attached a SECOND file (a PCOR - Preliminary Change of Ownership Report), pass it as pcorFilePath - this adds it as a Supporting Document (selects 'PCOR' from the Supporting Documents dropdown, clicks Add, uploads the second file) before finishing. Username and password are saved in Settings
(simplifileUsername / simplifilePassword} - NEVER ask the user for credentials, they are loaded automatically from config.

Searches Reddit for a keyword/topic (optionally within a specific subreddft) and extracts a batch of posts (author, title, text snippet, subreddlt, stable
postldentifier/permalink) - all in ONE tool call, no manual navigate/scroll/read_page loop needed. Use this FIRST when asked to find and comment on Redd it posts. After getting results, compose a unique comment for each post you want to engage with, then call browser_agent's post_comment_safely (platform: "reddit") once per post using the postldentifier returned here - this tool does NOT post anything itself.

Publishes a brand-new text post to a specific subreddit on Reddit - NOT a comment on an existing post. Give it the subreddit name, finished title, and body text; it handles navigating to the submit page, filling the form, and submitting in one hardcoded tool call. Reddit requires a subreddft to post into. For commenting on an existing post found via reddit_search, use browser_agent's
post_comment_safely (platform: "reddit") instead - this tool does not comment on anything.

Searches Linkedln (or reads the home feed) for a keyword/topic and extracts a batch of posts {author, text, stable postldentifier/permalink) from the results/feed - all in ONE tool call, no manual navigate/scroll/read_page loop needed. Use this FIRST when asked to find and comment on Linked In posts. After getting results, compose a unique comment for each post you want to engage with, then call browser_agent's post_comment_safely (platform:
"linkedin") once per post using the postldentifier returned here - this tool does NOT post anything itself.

Publishes a brand-new top-level post on Linked In - NOT a comment on an existing post. Give it the finished post text and it handles navigating to Linkedln, opening the post composer, typing, and submitting in one hardcoded tool call. For commenting on an existing post found via linkedin_search, use browser_agent's
post_comment_safely (platform: "linkedin") instead - this tool does not comment on anything.

Finds Facebook posts worth commenting on, but ONLY from groups the logged-in account has actually joined - it never searches public/global Facebook content, so it will never surface a group you're not a member of. Loads
'Groups you've joined', optionally verifies a specific requested group is really in that list, then navigates into the group(s} and extracts posts (author, text snippet, stable identifier) in one batch. Use this FIRST when asked to find and comment on Facebook posts. After getting results, compose a unique comment for each post you want to engage with, then call browser_agent's posLcomment_safely once per post (using the postldentifier returned here) to actually post- this tool does NOT post comments itself.

Publishes a brand-new top-level post on the user's own Facebook page/profile {the 'What's on your mind?' composer) - NOT a comment/reply on someone else's post. Give ft the finished post text and it handles navigating to Facebook, opening the composer, typing, and submitting in one hardcoded tool call. For commenting on an existing post found via
facebook_search, use browser_agent's
post_comment_safely instead - this tool does not comment on anything.

Renders a professional vertical (108Qx1920, TikTok/Reels format) promotion video from an image plus marketing copy - animated scenes (hook, cover reveal with drop shadow, selling-line cards, optional quote card, call-to-action) with slow zooms and crossfades, optional background music. Compose the COPY yourself: a punchy hook (under 8 words), 2-4 short selling lines, optionally a quote from the/reviews, and a CTA