Kai Write logo

Kai Write

Community
cgallic
kai-write

Write a single piece of marketing content using the marketing harness frameworks and quality gates. Supports blog posts, LinkedIn articles, emails, cold outreach, press releases, ad copy, and TikTok scripts. Automatically loads the right framework, skill contract, and persona. Use when "write a blog post", "draft an email", "LinkedIn article", "cold outreach", "write ad copy", "press release", "TikTok script", or any single content creation request. For building a complete email system, use /kai-email-system instead.

Overview

Publishercgallic
Repositorykai-cmo-harness
Skill namekai-write
Stars
61
Forks
8
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by cgallic on GitHub. Read the source before you install it.

Installation

Install the Kai Write AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/cgallic/kai-cmo-harness/tree/main/harness/skills-v2/kai-write
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/cgallic/kai-cmo-harness.git /tmp/kai-cmo-harness
mkdir -p .claude/skills
cp -r /tmp/kai-cmo-harness/harness/skills-v2/kai-write .claude/skills/kai-write
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Kai Write in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Kai Write on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Kai Write is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Kai root note: knowledge/, harness/, and scripts/ paths in this skill live in the Kai install, not the user's project. Resolve them against the first ancestor directory of this SKILL.md that contains a knowledge/ folder (the Kai plugin root, ~/.claude/kai, or the kai-cmo-harness repo). MARKETING.md, memory/, and any output files live in the current project. If a referenced scripts/ command is not available in this install, say so, skip it, and continue with the file-based guidance — never fabricate its output.

Objective

One piece of content in the requested format, written against the framework and contract that govern that format and clean through every gate that applies to it — copy a human can approve and publish, not a draft that needs a second author. Format is the load-bearing decision: it selects the framework, the contract, the gate thresholds, the policy reference, and the ECO floor.

Done when

The work type varies by format — read it off the request, never default to one. Floors from harness/eco-floors.yaml.

FormatWork typeFloorContractFour U'sSEO lint
Blog post / SEO contentblog-postE5/C3/O3harness/skill-contracts/blog-post.yaml12/16required
LinkedIn articlesocial-postE5/C2/O3harness/skill-contracts/linkedin-article.yaml12/16skipped
TikTok scriptsocial-postE5/C2/O3harness/skill-contracts/social-post.yaml10/16skipped
Email (lifecycle)email-lifecycleE5/C3/O3harness/skill-contracts/email-lifecycle.yaml10/16skipped
Email (cold outreach)cold-emailE5/C4/O3harness/skill-contracts/cold-email.yaml10/16skipped
Meta / Google ad copypaid-ad-campaignE5/C4/O4harness/skill-contracts/meta-ads.yaml · harness/skill-contracts/google-ads.yaml10/16skipped
Press releasepress-releaseE5/C3/O3harness/skill-contracts/press-release.yaml12/16skipped
  • E — a file in workspace/ is E1. E3 is a named human approving the exact bytes; E5 is a non-actor reading the live artifact back and matching it to the approved copy.
  • C — C2 is every gate above passing at its threshold; C3 adds a named non-producer reading it end to end; C4 (cold email, paid ads) adds the legal or platform standard verified before send.
  • O — metric, source, baseline, threshold, window, and owner recorded before ship. A baseline written after ship is rejected. This skill produces the artifact and can reach C; it does not publish, and it never issues its own verdict.

Constraints

  • Load the framework and contract before writing. Never write from memory of the rules.
  • Ad copy loads the platform policy reference first. Confirm the platform before drafting. Full per-platform table: .claude/rules/architecture-and-memory.md. Ads pass platform TOS on top of the quality gates.
  • Cold email carries legal weight. harness/references/cold-email-rules.md + harness/references/advertising-compliance.md govern identity, opt-out, and consent basis. Hard requirements, not style notes.
  • Read MARKETING.md from the project root before asking anything. If missing, build it from CLAUDE.md, README.md, PROJECT.md, package.json, landing pages, and email/ad/analytics config using the template from /kai-email-system, then confirm the draft. Do not ask discovery questions the repo can answer.
  • Know three things before drafting: the persona (one of the 8 in knowledge/personas/_persona-index.md), the angle (the frame, not the topic), and the single CTA. With no brief supplied, build one from harness/brief-schema.md.
  • Banned words — instant reject: leverage, utilize, synergy, innovative, deep dive, circle back, touch base, moving forward, at the end of the day. AI slop — never: "In conclusion", "It's important to note", "In today's rapidly evolving", "This comprehensive guide", "Without further ado", "It's worth noting that".
  • X-not-Y binary clichés fail the gate in every form. The pattern reads as LinkedIn slop and slips past subjective scoring. Failing patterns, outside HTML comments and code fences: , not [a-z] · — not [a-z] · \bisn'?t [a-z][^.\n]+ — it'?s\b · \bIt'?s (a\|the\|an) [^.\n]+, not (a\|the\|an)\b · \bIf you [a-z][^,.]+, [a-z] · \bHere'?s the thing\b · \bLet that sink in\b · \bHot take\b · \bI'?ll be honest\b. Fix by collapsing to a single load-bearing claim or by parallel-positive contrast where both halves are affirmative.
  • One primary CTA, word count from the contract. SEO content also applies Algorithmic Authorship: conditions after the main clause ("Do X if Y"), instructions start with verbs, sentences under 20 words, bold the answer rather than query-matching terms.
  • Gate failures get named fixes, never rewrites. Max 2 retry cycles, each naming the specific failing rule. After 2 failures, deliver with failures listed and log the repeated diagnosis to memory/lessons.md.
  • Publishing is out of scope. Deliver with gate scores visible, saved to workspace/ when asked; nothing goes live without human approval.

Context

NeedLoad
Blog post / SEO contentknowledge/frameworks/content-copywriting/algorithmic-authorship.md + knowledge/frameworks/aeo-ai-search/aeo-ai-search-playbook-2026.md · knowledge/checklists/content-checklist.md · knowledge/checklists/seo-checklist.md
LinkedIn articleknowledge/channels/linkedin-articles.md
Email, lifecycle or coldknowledge/channels/email-lifecycle.md (+ harness/references/cold-email-rules.md for cold)
Meta adsknowledge/channels/meta-advertising.md + harness/references/meta-ads-rules.md · knowledge/checklists/meta-advertising-checklist.md
Google adsknowledge/channels/paid-acquisition.md + harness/references/google-ads-policy-reference.md · knowledge/checklists/paid-acquisition-checklist.md
Press release · TikTok scriptknowledge/channels/press-releases.md · knowledge/checklists/pr-checklist.md — knowledge/channels/tiktok-algorithm.md · knowledge/checklists/tiktok-checklist.md
Persona · brief · product contextknowledge/personas/_persona-index.md · harness/brief-schema.md · MARKETING.md (project root)

Gates: python scripts/quality_gates/four_us_score.py <file> · banned_word_check.py <file> · seo_lint.py <file> (SEO only). If the project has .claude/hooks/voice-gate.py, it fires PostToolUse on Edit/Write and catches voice violations at draft time; the gate run still applies the patterns above.

Escalate when

  • The format is ambiguous and the wrong guess would change the contract, policy reference, or gate threshold.
  • Ad copy is requested for a platform with no policy reference in harness/references/, or cold outreach where the consent basis or suppression list is unknown.
  • A claim the copy depends on cannot be sourced — a quantitative claim with no source does not ship.
  • The piece still fails a gate after two targeted fixes.
  • The request is a full email system rather than one piece — that is /kai-email-system.

Frequently asked questions

What does the Kai Write AI skill do?

Write a single piece of marketing content using the marketing harness frameworks and quality gates. Supports blog posts, LinkedIn articles, emails, cold outreach, press releases, ad copy, and TikTok scripts. Automatically loads the right framework, skill contract, and persona. Use when "write a blog post", "draft an email", "LinkedIn article", "cold outreach", "write ad copy", "press release", "TikTok script", or any single content creation request. For building a complete email system, use /kai-email-system instead.

Why use Kai Write on TypingMind?

Because you install it once and use it with any model. Kai Write is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Kai Write in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/cgallic/kai-cmo-harness/tree/main/harness/skills-v2/kai-write. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Kai Write?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Kai Write?

As many as you like. As long as a model supports skills, you can use Kai Write with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Kai Write AI skill free?

It is published on GitHub by cgallic. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇