kevinzai/commander/ccc-marketing

/ccc-marketing — marketing work, routed

CC Commander · ccc-marketing · one router over the marketing pods, a shared context file, and scoring tools that need no installs.

Each sub-skill lives in its own directory beside this file. Read ${CLAUDE_SKILL_DIR}/<sub-skill>/SKILL.md (on Codex, the same path relative to this file) and follow it. marketing-ops is the in-domain router: its routing matrix, cross-domain hand-offs and campaign sequences are canonical, so this file does not repeat them. The user's selected model is preserved.

Codex: ask each picker as a plain-text question; to start headless, run codex exec "Read skills/ccc-marketing/marketing-ops/SKILL.md, then <task>" from the repository root.

Sub-skills

Python tools by sub-skill, with usage: references/tools.md.

When to use / when not to

• Use for content plans and production, page and funnel conversion, SEO and AI-search visibility, email and paid campaigns, referral and retention programs, competitor pages, pricing and launches.
• Use ccc-seo for a site-wide technical SEO workflow, ccc-design for the visual build of a page, ccc-data for analytics pipelines and dashboards.

What happens

1. Load context. Run ls .agents/marketing-context.md marketing-context.md .claude/product-marketing-context.md 2>/dev/null. Read the first file found. When none exists and the task is more than a one-off edit, ask with AskUserQuestion (header "Context"): 🟢 Create it first with marketing-context (one guided interview, reused by every sub-skill) · Skip and use only what is in this conversation.

1. Route. An argument or a clear intent picks the sub-skill from the table above, using the routing matrix in marketing-ops/SKILL.md to separate near neighbours (copywriting vs copy-editing, email-sequence vs cold-email, page-cro vs form-cro). A multi-skill campaign follows the matching sequence in marketing-ops/SKILL.md. Otherwise ask with AskUserQuestion (header "Marketing"), recommended first:

• 🟢 Content — plan, write, edit or humanize content (content-strategy, content-production, copywriting)
• SEO and AI search — audits, AI citation, schema, site structure (seo-audit, ai-seo)
• Conversion — pages, forms, signup, onboarding, popups, paywalls (the CRO pod)
• Something else — a follow-up picker with channels, growth, intelligence, and pricing and launch

1. Load and run. Read ${CLAUDE_SKILL_DIR}/<sub-skill>/SKILL.md and follow it. Write deliverables to files the user can review (for example marketing/<topic>.md), not only to chat.

1. Score the output when a tool exists. Run the matching script from references/tools.md, for example:

``bash python3 "${CLAUDE_SKILL_DIR}/content-production/scripts/content_scorer.py" marketing/post.md python3 "${CLAUDE_SKILL_DIR}/content-humanizer/scripts/humanizer_scorer.py" marketing/post.md ``

Report the score it prints. When a score is below the sub-skill's bar, revise and re-run once before reporting.

Done when

• Routing only: the sub-skill file was read in this turn and the user knows which one ran.
• A deliverable was written: test -s <file> && echo ok printed ok, and any scoring script that applies ran in this turn with its result reported.
• A context file was created: python3 "${CLAUDE_SKILL_DIR}/marketing-context/scripts/context_validator.py" <file> ran and its score is reported.

Guardrails

• Never post, send, schedule or publish content, email a prospect or influencer, or change ad spend without authorization in this conversation; ask with a picker at that step.
• Never invent statistics, testimonials, reviews, customer names or competitor claims; mark anything unverified for the user to confirm.
• Never put customer personal data into prompts, scripts or committed files.
• Never route new work to content-creator; it is deprecated.

See also

• ccc-seo — site-wide SEO workflow
• ccc-design — landing page and visual build
• ccc-data — analytics pipelines and dashboards