Kai is a marketing operating system that runs inside Claude Code.
It uses Claude Code's operator experience (skills, subagents, hooks, memory) and adds marketing-specific infrastructure:
- Business profiling with archetype detection and overlay inference
- Audit engines that score businesses across 8 categories with real algorithms
- Proposal ranking with 5-factor weighted scoring, dependency resolution, and capacity pruning
- Approval workflows with risk-based auto-approval and state machine enforcement
- Compliance checking with 100+ rules and regex pattern matching
- Quality gates (Four U's scoring, banned word detection, SEO lint)
- 168 marketing knowledge files (41 playbooks, 27 frameworks, 24 checklists, 12 ad platform policies)
| Layer | Status | Notes |
|---|---|---|
| Runtime models + persistence | Built | Atomic writes, thread-safe, audit log |
| Audit engines (8 categories) | Built | Real scoring algorithms, 239 tests passing |
| Proposal ranking + bundling | Built | 5-factor weighted, dependency-aware |
| Approval + action lifecycle | Built | State machine enforced, tested |
| Compliance engine | Built | 100+ rules, 50+ regex patterns |
| Content pipeline + quality gates | Built | LLM-backed via scripts/content/engine.py |
| Knowledge base | Built | 168 files, production-grade marketing intelligence |
| Connector business logic | Partial | Request construction complete, HTTP transport not wired |
| Creative module | Partial | Generates briefs/templates, not final content |
| Watchers/monitoring | Partial | Threshold logic real, data feeds not connected |
| Agent loop / subagent orchestration | Planned | Scaffolding only |
| Remote automation scheduling | Planned | Defined in module manifests, not triggered |
For the full assessment, see docs/superpowers/specs/2026-04-03-system-current-state-report.md.
Option 1 — One-liner (paste into your terminal):
curl -fsSL https://raw.githubusercontent.com/cgallic/kai-cmo-harness/main/install.sh | bash
Option 2 — Manual (paste into Claude Code):
git clone https://github.com/cgallic/kai-cmo-harness.git /tmp/kai-install && cp -r /tmp/kai-install/harness/skills/kai* ~/.claude/skills/ && rm -rf /tmp/kai-install && echo "Installed! Type /kai to start."
First run: Type /kai-start for guided setup. It reads your codebase, creates MARKETING.md, and recommends your first command.
Or type /kai to see all 31 commands.
The clone layer. This is the Claude Code-style product shell:
- workspace profile
- brand memory
- module activation
- local runs
- remote runs
The differentiation layer. This is the marketing-specific operating system:
- quality gates
- archetype-specific workflows
- campaign and content orchestration
- approvals
- learning loop
These work today through Claude Code's skill system. They load marketing knowledge + framework instructions and let the LLM do the work:
"Write all my emails" → /kai-email-system
Maps every email your product needs, writes them all, checks quality, outputs ready for Loops.
"Create my ad campaign" → /kai-ad-campaign
Plans ads across Meta, Google, LinkedIn, TikTok with policy compliance.
"Plan my content for the month" → /kai-content-calendar
Generates a month of content mapped to keywords, personas, and business goals.
"Audit my marketing" → /kai-audit
Runs checklists across SEO, content, email, ads, social, CRO. Health score + prioritized fixes.
"Write a landing page" → /kai-landing-page
Full landing page copy using conversion psychology frameworks.
These work through the Python runtime with real algorithms:
- Business profiling: Load a profile, detect archetype, infer overlays
- Scored audits: 8 audit engines with severity-weighted scoring across 40+ checks
- Proposal ranking: 5-factor weighted scoring with dedup, dependency resolution, capacity pruning
- Approval workflows: Risk-based auto-approval, state machine, batch review
- Compliance checks: 100+ rules with regex pattern matching, violation tracking
- Run persistence: Atomic JSON writes with lineage tracking and audit log
First time you run any /kai command in a project, it:
- Reads your codebase — CLAUDE.md, README, package.json, routes, schemas, whatever exists
- Builds a runtime profile — workspace, brand, archetype/module defaults, proof points, channels
- Creates
MARKETING.mdas a readable export — ICP, personas, brand voice, landscape, and operating defaults - Does the work — writes, audits, plans, scores, and routes outputs through the gate
Every command after that should run from the same workspace/brand/module model. MARKETING.md remains useful for humans, but the runtime contract is moving into code.
Kai Marketing OS
-> outcome engine, quality gate, approvals, archetype modules
Kai Runtime
-> skills, subagents, hooks, memory, plugins, local/remote runs
Implementation
-> scripts/content, scripts/quality, gateway, agent, knowledge
The new canonical runtime models live in kai/runtime/.
These commands work through Claude Code's skill system. They load knowledge files and framework instructions for the LLM. They do not invoke the kai/ Python runtime directly.
| Command | What you get |
|---|---|
/kai-write |
One piece of content — blog, email, LinkedIn, ad, press release |
/kai-landing-page |
Complete landing page copy |
/kai-email-system |
Every email your product needs |
/kai-ad-campaign |
Full ad campaign across platforms |
/kai-content-calendar |
Month of planned + written content |
/kai-social |
Week/month of social posts for all platforms |
/kai-video |
Video scripts for TikTok, YouTube, Reels |
/kai-cold-outreach |
Cold email sequences |
/kai-newsletter |
Newsletter editions |
/kai-case-study |
Customer success stories |
/kai-repurpose |
1 piece → 15-25 pieces across platforms |
/kai-launch |
Full product launch — emails + ads + PR + content + social |
/kai-retarget |
Retargeting campaign setup |
/kai-influencer |
Influencer marketing campaigns |
/kai-webinar |
Webinar/event marketing |
/kai-podcast |
Podcast launch or guest strategy |
/kai-abm |
Account-based marketing for enterprise |
/kai-partnership |
Co-marketing campaigns |
| Command | What you get |
|---|---|
/kai-gate |
Quality score on any content |
/kai-audit |
Full marketing audit with health scores |
/kai-seo-audit |
Technical SEO audit with fix list |
/kai-cro |
Conversion rate audit for any page |
| Command | What you get |
|---|---|
/kai-brief |
Content brief before writing |
/kai-growth-plan |
Marketing plan for your stage |
/kai-brand |
Brand positioning + messaging framework |
/kai-budget |
Marketing budget planning |
/kai-retention |
Customer retention system |
| Command | What you get |
|---|---|
/kai-competitors |
Competitive teardown + sales battlecards |
/kai-surround-sound |
Get mentioned in AI answers |
/kai-analytics |
Analytics + attribution setup |
| Command | What you get |
|---|---|
/kai |
See all commands, organized by what you need |
Just launched? → /kai-growth-plan first, then /kai-email-system + /kai-landing-page
Need content? → /kai-content-calendar for the plan, /kai-write for individual pieces
Running ads? → /kai-ad-campaign handles the copy, /kai-cro audits your landing page
Want everything? → /kai-launch orchestrates all of the above into one coordinated campaign
153 marketing knowledge files that power the commands:
- 41 playbooks (growth loops, CRO, pricing, competitive intel, content repurposing...)
- 27 frameworks (SEO rules, persuasion psychology, copywriting formulas, AI search optimization)
- 24 checklists (technical SEO, ad launch, content quality, email, social media audit...)
- 17 channel guides (every major platform)
- 8 audience personas
- 12 ad platform policies (Google, Meta, TikTok, LinkedIn, and 6 more — 7,600+ lines of TOS so your ads don't get rejected)
Every piece of content gets quality-checked before it ships:
- Four U's score — is it Unique, Useful, Ultra-specific, Urgent?
- Banned word check — no "leverage", "synergy", or AI slop
- Platform compliance — character limits, policy rules, format requirements
- Claude Code (CLI, desktop app, or VS Code extension)
- That's it
MIT — use it however you want.