Skill

web-audit

CRO + competitive audit of any URL. Bulleted recommendations against the web-audit-checklist.md heuristics. Reads about-me/business-brain.md (for ICP context). Uses Firecrawl + Playwright (for login-walled pages). Read-only on the web — never submits forms or clicks destructive actions. Lands in reference/research/audits/<domain>-<date>.md.

npx claudepluginhub automatedmarketer/cowork-research --plugin cowork-research

Invocation

How this skill is triggered — by the user, by Claude, or both

Slash command

/cowork-research:web-audit

User invocable

Model invocable

Inline context

Default effort

Context Preview

The summary Claude sees in its skill listing — used to decide when to auto-load this skill

CRO + competitive audit of any URL. Bulleted recommendations.

Supporting Files

templates/web-audit-checklist.md

SKILL.md

52 lines · ~546 tokens

Similar Skills

cache-components

139.5k

Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.

cache-components

pdf

131.6k

Processes PDFs: extracts text/tables/images, merges/splits/rotates pages, adds watermarks, creates/fills forms, encrypts/decrypts, OCRs scans. Activates on PDF mentions or output requests.

11 files

document-skills

Stats

LanguageShell

Parent stars0

MaintenanceExcellent

Last CommitMay 14, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Stats

Actions

Help us improve

Share bugs, ideas, or general feedback.

web-audit

/web-audit — v0.1.0

CRO + competitive audit of any URL. Bulleted recommendations.

Inputs

Required: URL

Read on every run:

about-me/business-brain.md (for ICP context — audits should reflect THE USER's customer, not generic best practices)
templates/web-audit-checklist.md (the heuristics — user can edit)
projects/research/research-preferences.md (audit style: cro-only / brand-only / both)

Logic

Read input files

Fetch the URL via Firecrawl. If Firecrawl reports "blocked / login-walled," fall back to Playwright (with user approval — Playwright requires a browser session)

Walk the checklist section by section. For each item, return PASS / FAIL / N/A with a 1-sentence reason.

Group failures by severity (critical / important / minor)

Plan-then-approve: show user the plan ("I'll audit against [checklist sections X, Y, Z]. Output: bulleted recommendations grouped by severity. Land in reference/research/audits/-.md. OK?")

On user approval, write the audit

Append one-line entry to reference/research/_index.md

Append one-line entry to projects/research/memory.md

Output

reference/research/audits/<domain>-<date>.md (the audit, structured as bulleted list)

Hard rules

READ-ONLY on the web — never submit forms, never click destructive actions

Playwright session NEVER persists credentials

Plan-then-approve before any write

Cite the URL accessed-on date in the output footer

Voice for the audit itself

Direct, actionable, scannable. Each recommendation is one bullet. No essays. The user wants to know what to fix and why, fast.

/web-audit — v0.1.0

CRO + competitive audit of any URL. Bulleted recommendations.

Inputs

Required: URL
Read on every run:
- about-me/business-brain.md (for ICP context — audits should reflect THE USER's customer, not generic best practices)
- templates/web-audit-checklist.md (the heuristics — user can edit)
- projects/research/research-preferences.md (audit style: cro-only / brand-only / both)

Logic

Read input files
Fetch the URL via Firecrawl. If Firecrawl reports "blocked / login-walled," fall back to Playwright (with user approval — Playwright requires a browser session)
Walk the checklist section by section. For each item, return PASS / FAIL / N/A with a 1-sentence reason.
Group failures by severity (critical / important / minor)
Plan-then-approve: show user the plan ("I'll audit against [checklist sections X, Y, Z]. Output: bulleted recommendations grouped by severity. Land in reference/research/audits/-.md. OK?")
On user approval, write the audit
Append one-line entry to reference/research/_index.md
Append one-line entry to projects/research/memory.md

Output

reference/research/audits/<domain>-<date>.md (the audit, structured as bulleted list)

Hard rules

READ-ONLY on the web — never submit forms, never click destructive actions
Playwright session NEVER persists credentials
Plan-then-approve before any write
Cite the URL accessed-on date in the output footer

Voice for the audit itself

Direct, actionable, scannable. Each recommendation is one bullet. No essays. The user wants to know what to fix and why, fast.