Trusted — Risk Score 5/100
Last scan:1 day ago Rescan
5 /100
nano-banana2
Image generation skill using kexiangai.com API (imgEditNB2 endpoint) for text-to-image and image-to-image generation
Legitimate image generation skill with clear documentation, declared network access to a single external API, optional local key storage with proper file permissions (chmod 600), and standard shell/python execution for API calls.
Skill Namenano-banana2
Duration41.3s
Enginepi
Safe to install
Skill is safe to use. Ensure users only enable --use-local-key after reviewing the local credential storage behavior. No action required.
ResourceDeclaredInferredStatusEvidence
Network READ READ ✓ Aligned SKILL.md declares network access to https://agent.mathmind.cn/minimalist/api/img…
Filesystem NONE READ ✓ Aligned SKILL.md documents ~/.config/nano-banana2/.env access only with --use-local-key …
Shell READ READ ✓ Aligned SKILL.md command templates show bash execution; scripts/generate.sh executes cur…
1 findings
🔗
Medium External URL 外部 URL
https://agent.mathmind.cn/minimalist/api/imgEditNB2
SKILL.md:10

File Tree

4 files · 14.0 KB · 474 lines
Markdown 2f · 323L Shell 2f · 151L
├─ 📁 references
│ └─ 📝 api-guide.md Markdown 56L · 949 B
├─ 📁 scripts
│ ├─ 🔧 generate.sh Shell 118L · 3.0 KB
│ └─ 🔧 set_key.sh Shell 33L · 620 B
└─ 📝 SKILL.md Markdown 267L · 9.4 KB

Security Positives

✓ API endpoint clearly declared: https://agent.mathmind.cn/minimalist/api/imgEditNB2
✓ API key masked in logs per SKILL.md security requirements
✓ Parameter validation with safe fallbacks (aspectRatio → auto, imageSize → 1K)
✓ Anti-repeat call protection prevents duplicate API calls
✓ 10-minute timeout (600s) explicitly set to handle long generation times
✓ Local key storage sets chmod 600 on .env file for proper permissions
✓ set_key.sh reads key from TTY to avoid command-line history exposure
✓ Input sanitization using grep with regex to safely parse .env file
✓ No credential exfiltration or hidden data transmission observed
✓ No obfuscation, base64 execution, or suspicious patterns detected