Marketing Audit Orchestrator
Run `/market audit <url>` and five parallel subagents score any website across content, conversion, SEO, competition and strategy, producing a client-ready marketing audit with revenue estimates.
The skill only fetches user-supplied URLs via WebFetch and writes MARKETING-AUDIT.md locally; no credential access, no covert exfiltration, no destructive defaults, and data flow is largely transparent. Deducted for: no confirmation mechanism for arbitrary URL fetching, no rollback guidance, and the README endorsing curl|bash remote script execution.
Instructions are self-consistent with explicit scoring weights and output format, plus an error-handling section (unreachable URL, subagent failure, auth-gated sites, thin content). Deducted for: static review cannot verify parallel subagent execution or Python script key paths; failure feedback quality unverified.
Trigger condition is clear (/market audit <url>) with a business-type classification table and defined scenarios. Deducted for: undeclared capability boundaries (JS-heavy or anti-scraping sites), no Chinese-language support, and full dependence on WebFetch to overseas sites, raising mainland-China reachability concerns.
Documentation is well structured, MIT license is clear, uninstall path provided. Deducted for: no version number, no changelog, no known-limitations section, unclear maintenance responsibility and update path, and unstated dependency on external agents/ directory files.
Report structure and scoring methodology are concrete and actionable in principle, likely yielding a client-usable audit. Deducted for: README sample output is illustrative rather than real run evidence, and revenue-impact estimates rest on largely unverified assumptions; actual output quality cannot be confirmed statically.
Scoring weights and output format are auditable, but evidence is limited to author claims and illustrative demo output — no test suite, no CI evidence, no third-party corroboration; key claims like revenue lift are not independently verifiable.
- The installer uses curl|bash to execute remote code directly; prefer manual clone and review install.sh before installing.
- Core function depends on WebFetch of target sites, which may be unreachable from mainland-China networks for many overseas sites, and there is no Chinese-language output support.
- Revenue-impact estimates are formula-driven assumptions, for reference only — not commitments to clients.
- This is a static review; no code was executed, so actual report quality, subagent parallel behavior, and script correctness remain independently unverified.
- No versioning or changelog, so behavior may change across updates without traceability.
What does this skill do, and when should you use it?
This is the flagship skill of the AI Marketing Suite for Claude Code, triggered by `/market audit <url>`. It fetches a site's homepage plus up to five key interior pages, detects the business type from six categories, then launches five parallel subagents covering content, conversion, competitive positioning, technical SEO, and brand/growth strategy. Results are combined into a weighted 0-100 Marketing Score with a letter grade, plus quick wins, strategic recommendations, a competitor comparison table, and revenue impact estimates. The output is a structured MARKETING-AUDIT.md report designed to be delivered directly to clients. It suits freelancers and small agencies selling AI-powered marketing services.
1) Uses WebFetch to retrieve the homepage and up to five key pages (pricing, about, product, blog, contact); 2) classifies the business as SaaS, e-commerce, agency/services, local business, creator/course, or marketplace using detection signals; 3) runs five parallel subagents: market-content (copy and messaging), market-conversion (CRO and funnels), market-competitive (positioning), market-technical (SEO and page performance), and market-strategy (brand trust and growth); 4) computes a composite score with 25/20/20/15/10/10 weights and assigns an A-F grade; 5) sorts recommendations into quick wins (<1 week), strategic (1-4 weeks), and long-term (1-3 months), estimating monthly revenue lift via traffic × conversion improvement × average deal value; 6) writes a structured MARKETING-AUDIT.md and prints a condensed terminal summary with score bars and top priorities.
- An agency salesperson audits a prospect's website before a sales call to arrive with specific scores and a gap list.
- A freelancer runs the audit on their own landing page and fixes headlines, CTAs, and missing meta descriptions from the quick wins list.
- A SaaS founder identifies weak points in trial-to-paid conversion, getting pricing-page and signup-flow recommendations.
- A marketing consultant needs consistent, scored audit deliverables with revenue estimates across multiple client sites.
- A local business owner checks for local SEO gaps such as NAP consistency, Google Business Profile, and reviews.
What are this skill's strengths and limitations?
- A clear three-phase pipeline (discovery → parallel analysis → synthesis) with a fully specified output format, ready as a client deliverable.
- Scoring weights and the revenue-impact formula are explicit, making reports explainable and reproducible.
- Business-type detection tunes analysis focus across six categories instead of generic advice.
- Built-in degradation: it continues when a subagent fails and adapts scope for single-page sites.
- Cross-skill integration pulls in COMPETITOR-REPORT.md and BRAND-VOICE.md when present.
- Hard dependency on Claude Code's parallel subagent mechanism and WebFetch; not portable as-is to other platforms.
- Revenue impact figures come from a generic formula (traffic × conversion lift × deal value) with assumed inputs; actual lifts are unverified.
- Competitor comparison quality depends on the subagent's ability to identify competitors; the source gives no reliability evidence.
- No test suite or benchmark data in the source to independently validate scoring accuracy.
- Analysis is limited to publicly fetchable pages; gated content is only flagged for manual review.
How do you install this skill?
Install the full suite (this skill ships with it): curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/ai-marketing-claude/main/install.sh | bash, or manually: git clone https://github.com/zubair-trabzada/ai-marketing-claude.git && cd ai-marketing-claude && ./install.sh. Optional PDF support: pip install reportlab. Uninstall with ./uninstall.sh. The source does not document installing this single skill in isolation.
How do you use this skill?
In Claude Code, type /market audit <url>, e.g. /market audit https://calendly.com. The skill fetches, analyzes, scores, and saves MARKETING-AUDIT.md to the current directory. If COMPETITOR-REPORT.md or BRAND-VOICE.md exist in the directory, their findings are incorporated. On unreachable URLs or subagent failures, the skill reports the error or notes the gap and continues with what it can analyze.