More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Catch Bad Design Exports Before Delivery

    1

    Design files can look finished while the export has the wrong size, name, profile or editable text. This skill gives your agent a delivery gate and a measured report before anything leaves for a client or printer.

    $7.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Catch Bad Design Exports Before Delivery

    Catch Bad Design Exports Before Delivery

    Example session with this skill installed

    Example request

    Gate this job before I send it to the printer (fictional job). Harbor Coffee Roasters open-day flyer, A5 portrait (148 x 210 mm), 3 mm bleed, printed on a digital press. Brief is signed off by the owner on 3 Nov. Facts from the working files:

    • Hero photo beans.jpg is 1200 x 800 px, placed full width across the bleed (154 mm wide). Licence state: licensed.
    • Small print line "Open day Saturday 15 Nov, 10:00 to 16:00" in #8a8a8a on white, 8 pt.
    • Exported harbor_flyer_A5_v02.pdf is RGB, page size 154 x 216 mm, saved 13:05. The source file was last saved 14:20.
    • Copy file says the address is "12 Quay Road"; the flyer artwork says "12 Quay Rd".
    • QR code was scanned on a phone screen and opens the booking page.
    • File name pattern agreed: brand_item_size_vNN.pdf.
    • Read your context and instructions
    • Compiled the catch bad design
    • Generated the UI component

    Example output

    Delivery gate: Harbor open-day flyer (8 Nov 2026) Verdict: HOLD (5 FAIL)

    Only the lines the request gave facts for are judged below. Every other checklist line has no evidence yet and counts as NOT ASSESSED, which blocks SHIP until it is answered.

    PhaseItemResultEvidence
    2 CopyText matches copy fileFAILArtwork "12 Quay Rd"; copy file "12 Quay Road". Do not change either: ask the owner which is right, then make the artwork match the copy file
    2 AssetsImage resolution at final sizeFAIL1200 px / (154 mm / 25.4 mm per inch) = 1200 / 6.063 = 197.9 ppi, under 300 ppi for a handheld flyer. Needs at least 6.063 x 300 = 1819 px across the 154 mm width, or a smaller placement
    2 TypeContrast of small printFAIL#8a8a8a on white = 3.45:1, under 4.5:1; at 8 pt it is small text. #6b6b6b gives 5.33:1
    3 ExportColour spaceFAILPDF is RGB; the digital press asks for CMYK or its own profile (confirm with the printer which)
    3 ExportFreshnessFAILExport saved 13:05, source saved 14:20: the PDF is older than the source. Re-export, then open the new PDF
    3 ExportPage size with bleedPASS148 + 2 x 3 = 154 mm wide, 210 + 2 x 3 = 216 mm high; matches the PDF
    3 ExportFile namePASSharbor_flyer_A5_v02.pdf fits brand_item_size_vNN.pdf (bump to v03 after re-export)
    2 CopyQR codePASSScanned on a screen, opens the booking page. Also scan a printed proof before the press run
    1 BriefSign-offPASSOwner approval dated 3 Nov

    Open FAILs and owners: copy wording (owner), photo (designer, supply a 1819 px or wider original or shrink the placement), contrast (designer), colour space (designer after printer confirms), re-export (designer).
    Waived: none.
    Delivery note: not sent. Needs the owner's answer on "Road" versus "Rd".

    harbor_flyer_A5_v02.pdf

    TSX · React component

    Generated

    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

    Design files can look finished while the export has the wrong size, name, profile or editable text. This skill gives your agent a delivery gate and a measured report before anything leaves for a client or printer.

    The proof

    These are real lines from reports saved in the package.

    From sample/bad-output.txt:

    FAIL | pixel-size       | hero-art.svg | 800x800; expected 1600x1000 | FIX: Export at exactly 1600x1000 pixels.
    FAIL | colour-profile   | hero-final.png | no ICC profile | FIX: Embed an sRGB ICC profile in the PNG export.
    SUMMARY: 33/100 | 4 PASS | 8 FAIL | 12 checks
    

    From sample/fixed-output.txt:

    PASS | colour-profile   | campaign_hero_1600x1000.png | sRGB IEC61966-2.1
    PASS | svg-live-text    | campaign_hero_1600x1000.svg | Campaign headline
    SUMMARY: 100/100 | 12 PASS | 0 FAIL | 12 checks
    

    What you get

    • SKILL.md: the brief, production, export, feedback and sign-off gate.
    • scripts/delivery_check.py: the standard-library export checker.
    • sample/: bad and fixed hand-off folders with saved reports.
    • tests/: pytest coverage for metadata, dimensions, names, profiles, scoring and the command line.
    • example-task.md and example-output.md: a worked delivery review.
    • A clear SHIP or HOLD output contract for the checks the agent completes.

    How to use

    1. Write a JSON specification for the required formats, size, name pattern and byte limit.
    2. Put the exports in one hand-off folder.
    3. Run python scripts/delivery_check.py HANDOFF --spec spec.json.
    4. Fix each FAIL and rerun the same command.
    5. Complete the manual production and sign-off lines before delivery.

    Known limitations

    • The script measures file metadata and SVG structure. It does not render the artwork.
    • It cannot judge taste, hierarchy, copy accuracy, licence validity or print quality.
    • A high automated score does not override an unanswered manual check.
    • Printer and platform requirements still need to be confirmed for each job.

    FAQ

    Does it approve a design? No. It measures export facts and provides a manual gate. A person still judges the design.

    Which agents does it work with? It is a SKILL.md skill for Claude Code and OpenAI Codex. The script can also run by hand.

    What files can it inspect? It reads PNG, JPEG and SVG metadata. The specification is JSON.

    Does it use the network or read secrets? No. It reads the folder and specification you give it.

    Can it save a report? Yes. The optional --json argument writes the measured report to the path you choose.

    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.

    Cannot perform subjective aesthetic reviews. Requires manual spec.json configuration for automated checks. Script only validates technical metadata, not visual content.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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 5 days ago

    Software
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed5 days ago

    What's inside

    Frequently Asked Questions

    Two layers. A script measures the exported files: required formats, file names, byte limits, pixel size, aspect ratio, embedded sRGB profiles in PNG/JPEG and live SVG text. A checklist covers what a script cannot measure (brief, copy, licences, type, colour, layout), each line answered PASS, FAIL or N/A with evidence. One FAIL or unanswered line means HOLD.

    Yes. It checks the exported files, not the design software. The script checks PNG, JPEG and SVG exports. PDF and other print files are not read by the script; they go through the checklist.

    SKILL.md (the brief, production, export, feedback and sign-off gate), scripts/delivery_check.py, a bad and a fixed sample hand-off folder with their spec.json files and saved reports, 12 pytest tests, a worked example, README, TESTING.md and CHANGELOG.md.

    No. You need Python 3 (standard library only) and a short spec.json listing the formats, size, file-name pattern and byte limit. The agent runs the script and works through the checklist. pytest is only needed to run the tests.

    The script does not look at the design. Legibility, WCAG contrast, copy matching and layout are checklist lines the agent answers with evidence, such as a measured contrast ratio or a compared copy line. Taste stays with you.

    It is a SKILL.md skill, tested with Claude Code and OpenAI Codex. The script is standard-library Python 3, so you can also run it by hand on any hand-off folder.

    New in UX & UI Design