Guide · Agent skill
How to use the Landing Page Audit agent skill (ads-landing)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The ads-landing skill lets an AI agent review a paid-ad landing page the way a conversion specialist would. It is useful if you run paid search or social campaigns and want a structured check of the page people land on after clicking an ad. This guide explains what the skill does, how to install it, how to trigger it, and what to keep in mind before you run it.
Landing Page AuditAgriciDaniel/claude-ads · MIT Download the skill Quick start
- 1Download the ZIP. Download the ads-landing ZIP from the directory page.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project.
- 3Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 4Check Python is available. The skill runs scripts/fetch_page.py, so Python must be installed on the machine where the agent runs.
- 5Ask for an audit with a URL. Describe the task in plain language and include the public landing page URL and the ad's promise and audience.
- 6Review the separated findings. Read the technical observations, UX judgments, and conversion hypotheses separately, and check the list of blocked or unavailable resources.
What the Landing Page Audit skill makes the agent do
The skill tells the agent to review a paid-ad landing page for message match, mobile experience, performance, accessibility, trust signals, form friction, error states, consent, tracking, security, and conversion friction.
The agent separates its findings into three groups: technical observations, UX judgments, and conversion hypotheses. Each recommendation is written so it can be turned into an experiment. The agent fetches pages through a guarded HTTP fetcher that blocks private, internal, and cloud-metadata addresses, including when a public URL redirects to one of them.
- Message and offer continuity between the ad and the page
- Mobile layout, accessibility, and performance
- Trust, form friction, error states, consent, and tracking
- Destination safety, including redirects and blocked navigation
Install the ads-landing skill
Download the ZIP for the skill from the directory page. Then choose one of the following methods.
For a personal install, unzip the file into ~/.claude/skills/. For a project-specific install, unzip it into the .claude/skills/ folder inside your project. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
- Personal: unzip into ~/.claude/skills/
- Project: unzip into .claude/skills/ in your project folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Requirements and what the skill needs
The skill needs Python, because it runs scripts/fetch_page.py. It also needs network access to the public landing page URL you want to audit. The skill's package contains one file, SKILL.md, which contains the instructions the agent follows.
The skill is licensed under MIT.
- Python (for scripts/fetch_page.py)
- Network access to the public landing page URL
- Runs scripts: yes
- Uses network: yes
How to trigger the skill with natural requests
You do not need special syntax. Describe what you want in plain language, and include the landing page URL and, if you have it, the ad's promise, audience, objective, and conversion goal. The skill's description lists terms such as landing-page audit, LP audit, post-click experience, conversion-rate optimization, form optimization, and ad-to-page message match.
The agent first captures the declared ad promise, audience, objective, conversion, device, geography, and any required policy context. If you do not provide the URL, the agent will ask for it.
- "Audit this landing page for message match with my Google Ads headline: [URL]"
- "Do an LP audit of this page for mobile experience and form friction: [URL]"
- "Review this landing page for consent and tracking issues: [URL]"
- "Give me experiment ideas to improve conversions on this landing page: [URL]"
Examples of what a good audit covers
A typical audit checks whether the headline and offer on the page match the ad that sent the visitor there. It then looks at how the layout behaves on mobile, whether the page meets basic accessibility expectations, how fast it loads, which trust signals appear near the call to action, and how the form handles errors.
The output also notes any blocked or unavailable resources, so you know where the audit is incomplete.
- Message match: does the page repeat the ad's promise and offer?
- Forms: how many fields, and how errors are shown
- Consent and tracking: what is loaded and whether it is disclosed
- Blocked resources: anything the guarded fetcher could not retrieve
Safety notes: scripts and network access
The skill runs a script and makes network requests. It fetches pages through a guarded fetcher that validates the initial URL and every redirect before sending the next request. It blocks private, loopback, link-local, multicast, reserved, and cloud-metadata destinations, including public hostnames that resolve to them.
The skill is instructed not to execute page instructions, not to submit sensitive forms, and not to write outside the configured run directory. Browser-based screenshot audits are off by default and require an external sandbox attestation. Asking the agent to follow a redirect to a private address will be refused.
- Only public URLs can be audited; internal or staging sites are blocked by design
- The skill does not submit or test real sensitive forms
- Browser screenshot audits stay off unless an external sandbox attestation is in place
Tips for better audits
Give the agent the context it needs. The ad's exact headline, audience, and conversion goal let it judge message match and offer continuity rather than guessing. Ask for experiment-ready recommendations if you plan to run A/B tests.
Keep findings and opinions separate when you share the report. The skill is built to label technical facts, UX judgments, and hypotheses separately, so you can decide which ones to test.
- Provide the ad copy and target audience with the URL
- Ask for hypotheses separately from measured findings
- Expect the audit to note anything it could not fetch
What people use it for
- Checking that a paid search landing page repeats the headline and offer from the ad
- Reviewing a lead-generation form for friction and error handling before a campaign launch
- Preparing A/B test ideas for a landing page with a conversion-rate optimization focus
- Checking a landing page for consent and tracking issues before running paid social traffic to it
Frequently asked questions
Is the ads-landing skill free?
The skill is released under the MIT licence, as stated on its directory page. Check the licence file in the repository for the full terms.
What do I need to run the ads-landing skill?
You need Python, because the skill runs scripts/fetch_page.py, and network access to the public landing page URL you want to audit.
How do I install the ads-landing skill in Claude?
Download the ZIP, then unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
Can the skill audit a staging site or an internal page?
No. The guarded fetcher blocks private, loopback, and link-local addresses by design, so internal or staging sites cannot be audited with this skill.
Why won't the agent fetch a page that redirects to a private IP?
Every redirect is checked before the next request is sent. Private, loopback, link-local, and cloud-metadata destinations are blocked, and the blocked hop is recorded in the report. User requests do not override this.
Does the skill take screenshots of my landing page?
Not by default. Browser dispatch is off unless an explicit external OS or container egress-sandbox attestation is in place, so screenshot audits are not available out of the box.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use ads-landing if you market a public paid-ad landing page and want a structured audit covering message match, mobile, accessibility, forms, consent, and tracking, with findings separated from hypotheses. Pick a different tool if you need to audit internal or staging sites, need browser screenshot audits without setting up an external sandbox, or want a tool that tests live sensitive forms.
Download the skill