AI Marketing Suite Orchestrator (market)
Run website marketing audits, generate copy, and deliver client-ready reports from the Claude Code command line.
The orchestrator transparently declares command routing, parallel subagents and file outputs; no malicious instructions or covert exfiltration. But it does not address authorization for fetching target websites, user confirmation, rollback, or data-flow disclosure for WebFetch requests to external sites — deducted.
Command table, routing logic and scoring weights are internally consistent; quick mode explicitly avoids subagents. But key paths (parallel subagents, Python scripts) are declared only: no tests, no edge-case handling or failure-feedback description, and execution is unverifiable in static review — deducted.
Audience (entrepreneurs, agencies, solopreneurs) and 15 scenarios are clear; business-type detection is explicit. But no declared capability boundaries or non-fit ranges, and core function depends on WebFetch of websites and reportlab with no assessment of Chinese-language support or mainland-China reachability — deducted.
Well-layered structure (orchestrator/sub-skills/agents/scripts/templates), MIT license, uninstall path, stable naming. But no versioning, changelog, or known-limitation disclosure; maintenance responsibility only implicit in a personal repo — deducted.
Output standards are specific (actionable, prioritized, example-driven, client-ready) and scoring methodology is defined; architecture plausibly yields Markdown/PDF reports. But no representative output verified as directly usable, and the calendly example is promotional rather than evidence — deducted.
Scoring methodology, file structure and dependency list are auditable; install/uninstall paths documented. But sample outputs are marketing demos, no test suite or third-party execution evidence, and the advertised 69/100 audit is not reproducible — deducted.
- The skill fetches and analyzes arbitrary URLs with parallel subagents; confirm authorization and compliance for target sites, especially when auditing third-party websites for clients.
- Sample outputs in the README are promotional demos; actual report quality is unverified without execution. PDF support requires installing reportlab separately.
- Core function depends on fetching external websites; mainland-China reachability is unassessed and Chinese-language output is not promised.
- Publisher is unverified; no versioning or changelog, and no explicit maintenance/update commitment.
What does this skill do, and when should you use it?
This is the main orchestrator skill of the AI Marketing Suite repository, located at market/SKILL.md. It routes 15 /market-prefixed commands to their sub-skills; the flagship /market audit launches 5 parallel subagents that score a website across content, conversion, competitive, technical, and strategy dimensions for a weighted 0-100 Marketing Score. Detailed output is saved as Markdown files, with optional client-ready PDF reports. The skill leans heavily on Claude Code-specific mechanisms (subagents, slash commands) and targets entrepreneurs, agency builders, and solopreneurs.
Accepts /market <command> slash commands and routes them to sub-skills under skills/; /market audit launches five parallel subagents (market-content, market-conversion, market-competitive, market-technical, market-strategy) and computes a weighted Marketing Score (Content 25%, Conversion 20%, SEO 20%, Competitive 15%, Brand 10%, Growth 10%); detects business type first (SaaS, e-commerce, agency, local, creator, marketplace) to tune the analysis; uses WebFetch to fetch pages for a 60-second quick assessment (/market quick); writes results to files like MARKETING-AUDIT.md and COMPETITOR-REPORT.md, with /market report-pdf producing a PDF (requires reportlab).
- An agency seller runs /market audit on a prospect's site before a sales call to walk in with a scored report
- A solopreneur uses /market copy to optimize their own landing page, getting before/after rewrite examples
- A product owner generates full email sequences and a launch checklist via /market emails and /market launch
- A content creator researches competitor positioning with /market competitors and plans posts with /market social's 30-day calendar
- A consultant pairs /market proposal with prior audit results and delivers a polished PDF via /market report-pdf
What are this skill's strengths and limitations?
- Broad command coverage: from a 60-second snapshot to full audits, email sequences, ad copy, client proposals, and PDF reports
- Transparent scoring methodology: six weighted dimensions and a 0-100 scale are documented directly in the skill
- Explicit output standards: actionable, impact-ranked, with before/after examples and client-ready formatting
- Adjusts analysis focus by detected business type (SaaS, e-commerce, local business, etc.)
- Tightly coupled to Claude Code-specific mechanisms (subagent orchestration, slash commands, WebFetch); porting elsewhere requires rewriting
- No test suite or benchmark in the repo, so scoring reliability is independently unverified
- Analysis quality depends on fetchable page content; behavior on login-walled or heavy-JS sites is undocumented
- PDF output requires an extra pip install of reportlab; Python script dependencies are only enumerated in requirements.txt, not itemized in docs
How do you install this skill?
Install the full suite (including this skill): curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/ai-marketing-claude/main/install.sh | bash, or git clone the repo and run ./install.sh. For PDF reports, run pip install reportlab. The skill itself lives at market/SKILL.md. Uninstall with ./uninstall.sh.
How do you use this skill?
In Claude Code, type /market <command> <url>, e.g. /market audit https://calendly.com or /market quick https://example.com. Audits launch parallel subagents and write Markdown files to the current directory; /market quick prints a sub-30-line scorecard to the terminal only. Note: the source does not document installing this single skill independently of the suite.