Security audit
Security checks for vulnerabilities and agentic risk
ClawScan rates this skill benign with medium confidence: the findings mostly fit its stated image workflow, but the install uses an unpinned package and one referenced script was only partly analyzed.
The verdict is benign, not risk-free. Before installing, consider that the installer version is not pinned and that prompts or reference images may be sent to an external image service. Avoid submitting private artwork unless you are comfortable sharing it. Confidence is medium because the script analysis was partial.
Referenced artifact was not completely inspected
translated from Japanese), then scripts/slice.sh cuts the sheet into frames, keys out the background,
Referenced artifact was not completely inspected
scripts/slice.sh needs bash and ImageMagick 7 (magick).
Referenced artifact was not completely inspected
contact match from frame to frame." That step is `scripts/slice.sh … --align` (section 3).
Referenced artifact was not completely inspected
## 3. Slice it: scripts/slice.sh
Referenced artifact was not completely inspected
bash scripts/slice.sh knight_run.png 4 4 # frames + GIF + atlas + JSON
Referenced artifact was not completely inspected
bash scripts/slice.sh knight_run.png 4 4 --key white --align # white -> transparent, one ground line
Referenced artifact was not completely inspected
bash scripts/slice.sh knight_run.png 4 4 --frames 14 --fps 10 # only the first 14 cells
Referenced artifact was not completely inspected
bash scripts/slice.sh run_strip.png 9 1 --name run # 9 frames in one row
The installation command uses an unversioned npm package through npx, allowing package code that changes over time to execute with the user's privileges. A compromised future release could therefore compromise the installation environment; this is a conditional supply-chain exposure, not evidence of malicious code, and the referenced package is an installer rather than an MCP server.
The examples invite users to process a local character image with the hosted GPT Image/varg workflow. Although the README names the services, it does not warn that reference images and prompts are transmitted to an external provider, which matters for private or unreleased artwork.
The workflow instructs the agent to submit the sprite specification and character-image URL to varg's image API, with fal specified in the provider options. Although the service and cost are documented, the skill does not warn that user-provided text and reference images are shared with external services.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
Without a character image, drop `files` and describe the character in one paragraph at the top. Over
REST the same `input` goes to `POST https://api.varg.ai/v2/image` ([gpt-image-2-6](../gpt-image-2-6/)
section 6). Make one sheet per action (run, idle, attack); keep the character image and the spec text
identical between them.