- Home
- Skills
- UX & UI Design
- ux audit
Works with the AI Agents you already use
ux audit
Audit UI/UX and accessibility to generate severity-ranked finding reports and executable fix plans for developers.
$5
ux audit
Example session with this skill installed
Run an accessibility-focused audit on the components in src/components/forms/ and generate a fix plan for any P0 or P1 issues found. Focus on keyboard navigation and screen reader labels.
- Read your context and instructions
- Compiled the ux audit
- Generated the UI component
I have completed the accessibility audit for the requested form components. The analysis identified two P0 blockers and three P1 high-priority issues related to missing labels and focus traps.
UX Audit Summary
| ID | Severity | Component | Finding | Criterion |
|---|---|---|---|---|
| A-01 | P0 | ContactForm.tsx | Submit button unreachable via keyboard | WCAG 2.1.1 (A) |
| A-02 | P0 | SearchInput.tsx | Input lacks programmatic label | WCAG 4.1.2 (A) |
| A-03 | P1 | SelectMenu.tsx | Low contrast on chevron icon (2.1:1) | WCAG 1.4.3 (AA) |
| A-04 | P1 | Form.tsx | Error messages not announced by SR | WCAG 3.3.1 (A) |
FIX_PLAN.md
# Fix Plan: Accessibility Sprint
## Batch 1: Global Form Primitives
**Files:** `src/components/forms/SearchInput.tsx`, `src/components/forms/Form.tsx`
- **Task A-02:** Add `aria-label` to the search input field.
- *Verification:* Screen reader should announce "Search site" when focused.
- **Task A-04:** Wrap error message containers in `aria-live="polite"`.
- *Verification:* Trigger a validation error and confirm the SR reads the message immediately.
## Batch 2: Interaction & Contrast
**Files:** `src/components/forms/ContactForm.tsx`, `src/components/forms/SelectMenu.tsx`
- **Task A-01:** Remove `tabindex="-1"` from the submit button wrapper.
- *Verification:* Ensure button is reachable using only the Tab key.
- **Task A-03:** Update chevron color from `$slate-300` to `$slate-500` to meet 4.5:1 contrast.
- *Verification:* Test with color contrast analyzer against background.
Next steps
- Review the
AUDIT_REPORT.mdfor full evidence and WCAG mapping. - Run
python3 scripts/tickets_to_csv.pyto import these into your tracker. - Assign Batch 1 to the next development turn.
ContactForm.tsx
TSX · React component
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Interface debt slows down development and frustrates users. This skill provides a structured framework to audit web and mobile applications, turning subjective design feedback into a ranked, executable engineering roadmap. It identifies accessibility gaps, design system drift, and usability friction points by triangulating evidence across source code, live URLs, and design files.
What it does
- Triangulated inspection compares source code, screenshots, and live environments to find contradictions between intent and implementation.
- Accessibility verification maps interface elements against WCAG 2.2 A/AA success criteria to identify legal and usability risks.
- Severity ranking categorizes findings from P0 (Blockers) to P3 (Polish) based on actual user impact rather than subjective taste.
- Plan generation produces a sequenced
FIX_PLAN.mdthat groups tasks by file and dependency for efficient agent execution. - Ticket synchronization exports findings into tracker-neutral formats or CSVs ready for Jira, Linear, or GitHub Issues.
How it works
- Scope definition starts with a read-only recon and interview to lock an
AUDIT_BRIEF.mdbefore any evaluation begins. - Evidence gathering collects technical artifacts including component code, CSS tokens, and rendered DOM states.
- Execution runs the checklists for accessibility, visual consistency, and usability to generate the
UX_AUDIT.mdreport. - Export generates executable fix batches and optional issue tracker tickets via the provided Python scripts.
Frameworks & tools
Works with any web or mobile stack including React, Vue, Tailwind, and React Native. Includes a Python-based CLI tool for converting ticket directories into Linear/Jira CSVs or GitHub CLI scripts.
Why this beats prompting it yourself
Generic prompts often return flat lists of subjective observations that lack file paths or severity context. This skill enforces a "triangulation" logic that proves findings with evidence, ensuring every reported issue is actionable and grounded in specific code locations.
Use cases
- Auditing a component library for WCAG 2.2 compliance before a major release.
- Identifying design system drift where implementation has diverged from Figma tokens.
- Converting a vague "UI polish" request into a prioritized backlog of P0-P3 tickets.
- Performing a pre-launch sanity check on keyboard navigation and focus management.
Known limitations
This skill focuses on the interface layer and does not profile backend architecture, API performance, or server-side logic. Audit depth is limited by the evidence provided (e.g., screenshots cannot reveal ARIA labels).
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for UX & UI Design.
Focuses only on UI/UX and accessibility. Does not audit backend, APIs, or performance. Depth is limited by provided evidence (code vs. screenshots).
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 27 days ago
Permissions
What it does: Audits a web app, mobile app, or component library's UI/UX — accessibility (WCAG 2.2), visual consistency/design-system drift, and usability/flows — and produces a severity-ranked findings report plus a fix plan another agent can execute without re-deriving the analysis. How it works: Gates on an approved AUDIT_BRIEF.md before anything runs. Takes inventory of whatever evidence exists (screenshots, source code, a live URL, design files — contradictions between sources are the highest-value findings). Runs three checklists (accessibility first, since it has objective pass/fail criteria). Assigns severity on a P0–P3 ladder keyed to user impact, not personal taste, with frequency multiplying severity. Writes exactly two artifacts: UX_AUDIT.md (findings, each with Where/Evidence/Impact/Fix) and FIX_PLAN.md (batched, agent-executable work ordered token-fixes-first since they cascade). Optionally emits tracker-neutral tickets via scripts/tickets_to_csv.py. Guardrails: Never promotes taste to P0/P1 — a severity has to name the user who gets hurt. Four hard gates: no audit without an approved brief, no scope creep, no tickets/ without asking, and it never modifies source — it audits, it doesn't fix. Judgment calls (severity ties, aggregations, exclusions) get logged to DECISIONS.md with a "Reverse it" line each.