Skills skywork-design
📦
skywork-design
Content revision r1 High Risk 🌐 Network access📁 Filesystem access🔑 Env variables⚙️ External commands
Create Visual Assets with Skywork
Teams need fast visual assets without manual design setup. This skill uses Skywork image workflows to generate or edit images from clear prompts.
Supports: Claude Codex Code(CC)
Stop for confirmation before installing.
Review the plan and obtain explicit user consent before changing files.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "skywork-design" from https://skillstore.io/skills/skyworkai-skywork-design.md and its manifest at https://skillstore.io/api/skills/skyworkai-skywork-design/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "skywork-design". Create a square hero image for a stainless steel travel mug.
Expected outcome:
- A 4K product hero image with a pure white background and centered composition.
- A local image file and a shareable Skywork OSS URL.
- A short status message if generation fails or account benefits are insufficient.
Using "skywork-design". Edit this logo reference into a premium stationery mockup.
Expected outcome:
- A 3:4 stationery mockup that preserves the provided logo and brand colors.
- A concise note describing the saved image and remote image link.
Using "skywork-design". Create a tri-fold brochure for a wellness studio.
Expected outcome:
- An outside brochure face with three panels and clear fold structure.
- An inside face can follow after user approval of the visual direction.
Security Audit
High Riskv5 • 7/7/2026 Open versioned report
The skill performs legitimate image generation work, but it uses external Skywork APIs, uploads user prompts and reference images, and stores reusable authentication tokens in a hidden home-directory file. Most prose-based static hits are false positives, while the confirmed risks are concentrated in network transfer, token handling, Windows shell launch behavior, and local file writes.
11
Files scanned
1,166
Lines analyzed
23
Review items
0
False positives ignored
Confirmed security concerns (2)
High
Prompts and Reference Images Sent to External API
The image workflow sends user prompts, optional base64 reference images, and an authentication token to Skywork API endpoints. This is disclosed as core behavior, but it can expose sensitive visual assets or prompt content to a third party.
The code directly constructs token-bearing HTTP requests and includes prompt text plus base64 image data in request bodies. This is clear third-party data transfer, although it is expected for the skill.
Medium
Unvalidated Upgrade Link From Script Output
The instructions tell the agent to extract an upgrade URL from script output and present it to the user. Without domain validation, compromised output could cause a user-facing phishing or payment-redirection link.
The instruction explicitly forwards a URL parsed from logs, but the observed text does not show host validation. The risk depends on the trustworthiness of backend error output.
Capability review items (23)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
High
Python subprocess.Popen
subprocess.Popen(["start", url], shell=True, start_new_session=True,
The Windows branch launches the browser with shell=True and a URL influenced by configuration. Shell interpretation creates an injection risk if the URL source is compromised or misconfigured.
High
Hidden file in home directory
Token is stored in user home directory as ~/.skywork_token (JSON format).
The code stores an authentication token in a hidden home-directory file. Persistent plaintext credential storage is security-sensitive and lacks permission hardening in the shown implementation.
High
Hidden file in home directory
TOKEN_FILE = os.path.expanduser("~/.skywork_token")
The code stores an authentication token in a hidden home-directory file. Persistent plaintext credential storage is security-sensitive and lacks permission hardening in the shown implementation.
High
Hidden file access
Token is stored in user home directory as ~/.skywork_token (JSON format).
The code stores an authentication token in a hidden home-directory file. Persistent plaintext credential storage is security-sensitive and lacks permission hardening in the shown implementation.
High
Hidden file access
TOKEN_FILE = os.path.expanduser("~/.skywork_token")
The code stores an authentication token in a hidden home-directory file. Persistent plaintext credential storage is security-sensitive and lacks permission hardening in the shown implementation.
High
Python file write/append
with open(TOKEN_FILE, "w", encoding="utf-8") as f:
The script writes token data to the cache file without setting restrictive file permissions. A reusable credential can remain exposed to other local processes or backups.
High
Hidden file in home directory
2. Cached token file `~/.skywork_token` → validate via API, if valid, use it
The documentation directs use of a cached token file in the user home directory. That matches the script behavior and represents persistent credential storage.
High
Hidden file access
2. Cached token file `~/.skywork_token` → validate via API, if valid, use it
The documentation directs use of a cached token file in the user home directory. That matches the script behavior and represents persistent credential storage.
Medium
Python HTTP libraries
req = urllib.request.Request(
The script sends POST requests to the Skywork gateway with a token header and prompt or image payloads. This is a real third-party data transfer risk, even though it is expected behavior.
Medium
Python HTTP libraries
with urllib.request.urlopen(req, timeout=300) as resp:
The script sends POST requests to the Skywork gateway with a token header and prompt or image payloads. This is a real third-party data transfer risk, even though it is expected behavior.
Medium
Hardcoded URL
SKYWORK_GATEWAY_URL = os.environ.get("SKYWORK_GATEWAY_URL", "https://api-tools.skywork.ai/theme-gate
The gateway defaults to an external Skywork endpoint and can be overridden by SKYWORK_GATEWAY_URL. Later token-bearing requests use this base URL, so endpoint control is security relevant.
Medium
Python file write/append
with open(out_abs, "wb") as f:
The script writes downloaded remote image bytes to a caller-provided filename. This is expected output behavior but can overwrite or create files outside the intended workspace.
Medium
Python os file operations
os.makedirs(os.path.dirname(out_abs) or ".", exist_ok=True)
The script creates directories derived from the output path. This is legitimate for saving images, but it can modify arbitrary filesystem locations if the filename is unsafe.
Medium
Python environment access
SKYWORK_GATEWAY_URL = os.environ.get("SKYWORK_GATEWAY_URL", "https://api-tools.skywork.ai/theme-gate
SKYWORK_GATEWAY_URL controls the external endpoint used for later token-bearing API calls. This environment override could redirect prompts, images, and credentials.
Medium
Hardcoded URL
API_BASE = os.environ.get("SKYWORK_API_BASE", "https://api.skywork.ai")
The API base defaults to Skywork and can be overridden by environment. Token validation and login-key requests use this endpoint, making the network destination security relevant.
Medium
Hardcoded URL
LOGIN_BASE_URL = os.environ.get("SKYWORK_WEB_BASE", "https://skywork.ai")
The login web base defaults to Skywork and can be overridden by environment. The resulting URL is shown to users and opened in a browser, so redirect risk exists.
Medium
Python environment access
API_BASE = os.environ.get("SKYWORK_API_BASE", "https://api.skywork.ai")
SKYWORK_API_BASE can redirect authentication traffic. Because token checks use this endpoint, the environment variable is security-sensitive.
Medium
Python environment access
LOGIN_BASE_URL = os.environ.get("SKYWORK_WEB_BASE", "https://skywork.ai")
SKYWORK_WEB_BASE controls the login URL shown to users and opened in a browser. A malicious value could redirect users to an unexpected authentication page.
Low
Python HTTP libraries
import urllib.request
The HTTP library import is used later for real outbound requests. Network access is central to the skill and can transmit user prompts or image data.
Low
Python HTTP libraries
req = urllib.request.Request(file_url, method="GET")
The script downloads a result image from a URL returned by the API. This is expected, but it still retrieves remote content and writes it locally.
Low
Python HTTP libraries
with urllib.request.urlopen(req, timeout=120) as r:
The script downloads a result image from a URL returned by the API. This is expected, but it still retrieves remote content and writes it locally.
Low
Python environment access
env_token = os.environ.get("SKYBOT_TOKEN", "")
The script reads SKYBOT_TOKEN from the environment as a credential source. This is an expected auth pattern, but it is still access to secret material.
Low
Python environment access
env_token = os.environ.get("SKYBOT_TOKEN", "")
The script reads SKYBOT_TOKEN from the environment as a credential source. This is an expected auth pattern, but it is still access to secret material.
Risk Factors
🌐 Network access (9)
📁 Filesystem access (9)
🔑 Env variables (6)
⚙️ External commands (30)
scripts/skywork_auth.py:87 scripts/skywork_auth.py:90 scripts/skywork_auth.py:93 SKILL.md:22-25 SKILL.md:25-28 SKILL.md:28-29 SKILL.md:29-32 SKILL.md:32-41 SKILL.md:41-43 SKILL.md:43-46 SKILL.md:46-48 SKILL.md:48-51 SKILL.md:51-53 SKILL.md:53-59 SKILL.md:59-60 SKILL.md:60 SKILL.md:63-75 SKILL.md:75 SKILL.md:85-89 SKILL.md:89-92 SKILL.md:92-93 SKILL.md:93-94 SKILL.md:94-98 SKILL.md:98-99 SKILL.md:99-103 SKILL.md:103 SKILL.md:104-105 SKILL.md:105-106 SKILL.md:106-113 SKILL.md:113-114
Audited by: codex View Audit History →
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/skyworkai-skywork-design/audits/5?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/skyworkai-skywork-design?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/skyworkai-skywork-design?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/skyworkai-skywork-design/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/skyworkai-skywork-design.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
SkyworkAI. (2026). skywork-design security audit report (audit version 5) [Author version unspecified]. Skillstore. https://skillstore.io/skills/skyworkai-skywork-design/audits/5BibTeX citation
@techreport{skyworkai-skyworkai-skywork-design-2026,
author = {SkyworkAI},
title = {skywork-design security audit report (audit version 5)},
institution = {Skillstore},
year = {2026},
number = {5},
url = {https://skillstore.io/skills/skyworkai-skywork-design/audits/5},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "skywork-design security audit report (audit version 5)"
version: "unspecified"
type: report
authors:
- name: "SkyworkAI"
date-released: "2026-07-07"
url: "https://skillstore.io/skills/skyworkai-skywork-design/audits/5"
identifiers:
- type: other
value: "skillstore:skyworkai-skywork-design:audit:5"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: Medium 45
Architecture
85
Maintainability
87
Content
70
Community
83
Spec Compliance
What You Can Build
Launch Product Image Sets
Create marketplace hero, lifestyle, and feature callout images with consistent sizing and visual style.
Develop Brand Concepts
Generate logos, palette cards, typography showcases, and stationery mockups from a shared brand direction.
Produce Campaign Visuals
Create social posts, posters, brochures, and story visuals in platform-specific aspect ratios.
Try These Prompts
Create a Simple Image
Create a 2K image of [subject] for [purpose]. Use a [style] look, [lighting] lighting, and a [aspect ratio] canvas.
Edit a Reference Image
Edit the attached image to [change]. Preserve [important elements], adjust [colors or layout], and keep the final image suitable for [channel].
Build a Brand Visual System
Create a visual identity for [brand] in [industry]. Start with a logo concept, then align palette, typography, and mockup style around [brand adjectives].
Generate an E-commerce Listing Set
Create a product image set for [product]. Include a white-background hero, lifestyle scene, feature callout image, and optional detail shots for [buyer concerns].
Best Practices
- Provide the target channel, audience, required aspect ratio, and final use before generating.
- Use reference images when consistency across a campaign or brand system matters.
- Review generated text, logos, product claims, and brand details before publishing.
Avoid
- Do not submit confidential product photos or private documents without user approval.
- Do not ask for many unrelated visual styles in one prompt.
- Do not use generated commercial assets without checking brand, trademark, and platform requirements.
Frequently Asked Questions
Does this skill work without a Skywork account?
No. It requires a valid Skywork token from the environment, cached login, or browser authentication.
Can it edit existing images?
Yes. It accepts one or more reference images and sends them with edit instructions to the Skywork API.
Which aspect ratios are supported?
It supports common ratios including square, portrait, landscape, mobile story, banner, and ultra-wide formats.
Where are generated images stored?
The script saves the image to the requested local filename and also prints the remote OSS URL.
Can it create full brand systems?
It can guide multi-step brand assets such as logos, palettes, typography samples, patterns, and stationery mockups.
What should users review before publishing?
Users should review image accuracy, readable text, product claims, brand compliance, and rights-sensitive content.