Skills package for AI agents

Give AI agents
real buyer context

Twelve skills that simulate how AI buying agents read your site, score it against your ICP, and ship a prioritized gap report your own agents can close.

View on GitHub

Claude Code · Codex · Cursor · OpenClaw · any Agent Skills runtime

How it works

Install. Anchor. Audit.

Twelve focused skills. One foundational profile every audit reads from. Reports land in ./reports/ — scored, prioritized, and ready for your repair agents to act on.

01 / Install

Add the skills package

One command installs all twelve skills under your agent's skill directory.

$ npx skills add RevensiAI/buyer-context

02 / Anchor

Define your buyer context

A 5-minute guided Q&A writes ./buyer-context.md — your ICP, positioning, USP, proof points. Every audit reads from this for alignment scoring.

/buyer-context

03 / Audit

Run a single surface or the whole site

Each report scores six dimensions, lists what's working, what's broken, and ranks fixes by impact × ease.

/full-audit example.com

The skills

Ten audits, one orchestrator.

All ten audits read from one foundational profile — buyer-context. Run a single audit for a focused pass, or let full-audit orchestrate the entire site in one go.

buyer-context

Foundation

Capture your ICP, positioning, and offer in one document. Every audit reads from this.

full-audit

Orchestrator

Run every audit in one pass and synthesize a cross-surface fix list, ranked by impact.

01

crawler-audit

Infrastructure

Check whether AI crawlers can actually reach your site — robots, sitemap, schema, llms.txt.

02

homepage-audit

Audit

Score the homepage for AI readability and return ranked fixes that change the verdict.

03

pricing-audit

Audit

Find pricing-page gaps that confuse buyers and stall agents looking for a quote.

04

about-audit

Audit

Audit your About / Team page for the trust signals AI uses to qualify a vendor.

05

comparison-audit

Audit

Surface weak spots in your /vs/ pages so AI doesn't paraphrase the competition.

06

features-audit

Audit

Pressure-test feature pages for clarity, proof, and the one job they actually do.

07

case-study-audit

Audit

Make your customer stories quotable by an AI — concrete numbers, named buyers, clean structure.

08

faq-audit

Audit

Tune the FAQ so AI agents can answer about you accurately, not approximately.

09

agent-page

Audit / Build

Audit your /for-ai-agents page — or generate one with the right JSON-LD if it doesn't exist.

10

competitor-audit

Intel

See a competitor's site exactly as AI describes it, then plan around the gaps you find.

See your site as the AI sees it.

View on GitHub