AgentSkillsCN

tweet

起草、迭代并润色推文。将屏幕截图嵌入终端窗口的 Chrome 浏览器中，方便分享。

SKILL.md
--- frontmatter
name: tweet
description: Draft, iterate, and polish tweets. Wraps screenshots in terminal window chrome for sharing.
user_invocable: true

Tweet Skill

Help the user draft, refine, and finalize tweets for posting. When screenshots are provided, wrap them in a macOS-style terminal window for polished sharing.

Input

$ARGUMENTS may contain:

  • •Raw tweet text to refine
  • •A file path to a screenshot to wrap
  • •Both text and screenshot path
  • •Nothing (interactive mode - ask what they want to tweet about)

Step 1: Analyze Input

Determine what the user provided:

  • •Text only → go to Step 2
  • •Screenshot only → go to Step 3
  • •Both → do Step 2 and Step 3 in parallel
  • •Nothing → ask what they want to tweet about

Step 2: Tweet Text Iteration

Initial Review

Read the draft and identify:

  • •Character count (280 limit, flag if over)
  • •Grammar/spelling issues
  • •Clarity problems
  • •Weak opening (first 5 words matter most in timeline)
  • •Missing context that makes the tweet confusing to outsiders
  • •Unnecessary filler words or hedging language

Generate Variations

Produce 3 variations plus the fixed original:

  1. •Fixed original - Same voice/intent, just fix obvious issues (grammar, spelling, awkward phrasing)
  2. •Punchy - Shorter, more direct, removes hedging. Frontload the hook.
  3. •Expanded - Adds context or a concrete example if the original is too vague
  4. •Hot take - Reframes as a stronger opinion or contrarian angle

For each variation show:

  • •The tweet text
  • •Character count
  • •What changed and why

Rules for Good Tweets

  • •No hashtags unless the user explicitly includes them
  • •No emojis unless the user's draft already uses them
  • •Don't start with "I" if possible
  • •Avoid corporate/marketing speak ("excited to announce", "thrilled to share")
  • •Match the user's voice - if they're casual, stay casual
  • •If referencing code/tech, be specific not vague
  • •Thread-style (1/n) only if user asks for it

Iterate

After showing variations, ask which direction they prefer or if they want to mix elements. Keep iterating until they're happy with the final version.

Step 3: Screenshot Wrapping

When the user provides a screenshot (file path or mentions wrapping an image):

3a. Read the Screenshot

Use the Read tool to view the screenshot and understand its content. Check:

  • •Is it a terminal/CLI screenshot? Use terminal window chrome.
  • •Is it a browser/web screenshot? Use browser window chrome.
  • •Dimensions - needed for the wrapper

3b. Generate and Run the Wrapper Script

Create a temporary Node.js script based on templates/wrap-screenshot.mjs. Customize:

  • •INPUT_PATH - the screenshot file path
  • •OUTPUT_PATH - same directory, append -twitter before extension (e.g., img.png → img-twitter.png)
  • •TITLE - contextual title for the window bar based on screenshot content:
    • •Terminal screenshots: claude — ~/project or similar
    • •Browser screenshots: URL or page title
    • •Generic: filename or brief description

The template outputs at 1600x900 (Twitter-optimal 16:9). The screenshot is scaled to fit inside the window and the window floats centered on the canvas. This prevents Twitter from cropping the image in timeline.

Run the script:

bash
node /path/to/scratchpad/wrap-screenshot.mjs

Dependencies (sharp, @resvg/resvg-js) should be available globally or in the project. If not, install to scratchpad:

bash
cd /path/to/scratchpad && npm init -y && npm install sharp @resvg/resvg-js

3c. Verify Output

Read the output image to verify it looks correct. Show the user the result.

Step 4: Final Output

Present the final tweet text (if applicable) and the wrapped screenshot path (if applicable). Offer to:

  • •Make further adjustments
  • •Generate additional variations
  • •Wrap more screenshots