agentleFS
Amine-MhiriAmine-Mhiri/creatorscompanion-plugin/skills/write-in-my-voice/SKILL.md

write-in-my-voice

Skill0 starsChanged 7 days agoMITView on GitHub

Draft a Reels or TikTok script, LinkedIn post, X post or thread, email newsletter, YouTube script or Instagram carousel in the user's own voice with CreatorsCompanion, from an idea, a link, or a post to remake. Use when the user asks to write, draft, rewrite, remake or repurpose social content or a newsletter. Checks the credit cost first and asks before spending credits, edits the draft on request, then offers to save it to the library with the hook the user picked.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. magnusrodseth/dotfiles★ 3same name
    2. jhouser224/claude-skills★ 0same name
    3. prateeks367/voicemoat-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 2 are distinctive, and 8 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 8

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the CreatorsCompanion tools to write drafts that sound like the user.

    line 8
  2. Very specific:

    If you do not know yet whether the account is set up, call get_started (free).

    line 12
  3. Very specific:

    Structure (optional): If the user asks for a specific structure, call list_architectures with the format and put the chosen keys in that …

    line 23
  4. Very specific:

    Call estimate_cost with the same formats (and source_url if there is one).

    line 41
  5. Very specific:

    Do not call generate_content until the user says yes: If they change the request, estimate again.

    line 45
  6. Distinctive:

    If the balance or today's limit does not cover it, say so plainly and suggest a smaller request.

    line 47
  7. Very specific:

    Call generate_content with the confirmed formats, the one source, and instructions (plus carousel_style when set).

    line 53
  8. Distinctive:

    If drafts come back, present them (step 4).

    line 56
  9. Very specific:

    If you get a job_id instead, the drafts are still being written.

    line 57
  10. Very specific:

    Ask whether to save any draft to their library, and which hook they want if there are several.

    line 86

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Write in my voice

Use the CreatorsCompanion tools to write drafts that sound like the user. The voice comes from their saved profile, voice examples and Creator Brain, so do not imitate it yourself: send the request to generate_content.

If you do not know yet whether the account is set up, call get_started (free). When the profile is empty, follow the setup skill first: one message with at most 3 questions, then save.

1. Pin down the request

Work out these, and ask one short question only if something is missing:

  • Formats. formats is a list with one entry per format, such as [{ "format": "tiktok" }, { "format": "linkedin", "variants": 2 }]. Format keys: tiktok (Reels and TikTok script), linkedin, twitter (X post or thread), newsletter, youtube, instagram_carousel. variants (1 to 4) asks for several drafts of one format.
  • Structure (optional). If the user asks for a specific structure, call list_architectures with the format and put the chosen keys in that entry's architectures. Otherwise leave it out and CreatorsCompanion picks one.
  • Carousel look (optional). For instagram_carousel, list_carousel_styles lists the styles; send the chosen slug as carousel_style. Leave it out for the house style.
  • Source, exactly one of:
    • idea: the topic in the user's words.
    • source_url: a Reel, TikTok, YouTube video, post or article to write from (adds a small link fee).
<!-- full-server -->
  • remix: a post the user already found, to remake in their voice with its structure kept and no link fee. Pass { "saved_key": ... } for a post from list_inspiration, { "job_id": ..., "post_key": ... } for a post from a find_viral_posts search, or { "post_id": ... } for a tracked creator's post from get_creator_report or get_feed.
<!-- /full-server -->
  • Extras. Angle, call to action, length or anything else goes in instructions.

2. Check the cost and ask

Call estimate_cost with the same formats (and source_url if there is one). Tell the user, in one or two lines, the total credits, their balance, and whether today's limit covers it. Then ask: "Shall I write it?"

Do not call generate_content until the user says yes. If they change the request, estimate again.

If the balance or today's limit does not cover it, say so plainly and suggest a smaller request. For how credits work, point to https://www.creatorscompanion.app/how-credits-work. Do not push the user to buy anything.

3. Write

Call generate_content with the confirmed formats, the one source, and instructions (plus carousel_style when set).

  • If drafts come back, present them (step 4).
  • If you get a job_id instead, the drafts are still being written. Tell the user, then call get_generation with that job_id until the status is done. Never call generate_content again for the same request while it runs.
  • An identical request made in the last few minutes is handed back at no cost (replayed). When the user wants new versions of the same request, confirm the cost again and set fresh: true.
  • Failed drafts give their credits back automatically. Mention it if it happens.
  • A tip line means the profile or voice is thin: relay it in one sentence and offer the setup skill.

4. Present

Some apps (Claude and ChatGPT) show the drafts as a card under the tool call, with hooks, on-screen text, caption, script and a Save button. When the card is shown, do not paste the drafts again in full: give a one or two line summary per draft and the cost. When there is no card, show each draft cleanly with its parts labeled (for a Reels or TikTok script: hooks, on-screen text, script and caption). Keep the wording as written; do not rewrite drafts unless asked. End with the cost and the credits left, as reported by the tool.

5. Edit on request (free)

  • Text changes to a draft: update_draft with the draft_id and only what changes: title, body, hooks, ctas, onscreen_hooks, caption (LinkedIn and X drafts have options; set option to edit one other than the first). Lists replace the whole list, in order.
  • Carousel slides: edit_carousel with the draft_id and slides (each by number, with the new headline, subheadline, body, cta_action or cta_reward), a new caption or a new carousel_style. It re-renders the slides; if it returns a job_id, use get_generation.
  • A draft the user does not want: discard_draft with its draft_id, after they confirm.

6. Offer to save (free)

Ask whether to save any draft to their library, and which hook they want if there are several. On yes, call save_to_library with the draft_id and hook (its number, 1 for the first, or its text). The library tags its content pillar and funnel stage on its own. To save text the user wrote in the chat instead, set title, text and platform (instagram_tiktok, linkedin, twitter, email, youtube or carousel). Changes to an item already saved go through update_library_item.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.