Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Pic Gen
v1.0.2AI 图片生成与提示词优化工具。支持通义万相、Banana (Flux)、DALL-E 等多模型。用于：根据用户简单描述生成高质量图片提示词、优化已有提示词、直接调用配置好的模型出图。触发场景：「生成图片」「画一个 XXX」「优化这个提示词」「设置 API key」「切换模型」。用户可直接发送密钥给 bot，自动...
⭐ 0· 107·1 current·1 all-time
byRoland Dickens@chenlong1314
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description match the code: scripts perform prompt optimization and call Qwen (DashScope), Banana/Flux, and DALL‑E APIs. Requesting multiple provider keys (DashScope, Banana, OpenAI) is coherent for a multi-backend image generator. However, the registry metadata declares no required environment variables or primary credential while the SKILL.md and scripts clearly expect DASHSCOPE_API_KEY, BANANA_API_KEY, and OPENAI_API_KEY (or config/models.yaml) — this mismatch is unexpected.
Instruction Scope
SKILL.md instructs the bot to ask users for API keys via chat and to write those keys into config/models.yaml. The scripts read/write that file and also read environment variables. Asking users to paste secrets into chat and then storing them unencrypted on disk extends the skill's scope into credential collection and persistent storage — a privacy risk that should be explicitly considered by users.
Install Mechanism
There is no install spec (instruction-only), and included Python scripts rely on standard libraries and the listed requirements.txt (requests, pyyaml, banana-dev). No external arbitrary downloads or obscure install URLs are present.
Credentials
The skill legitimately needs API keys for the image providers it supports. That said, the registry metadata does not list required env vars while the documentation and scripts use multiple env vars and a local config file. Requiring three different provider keys is proportionate to a multi‑backend image generator, but users should be aware those secrets will be stored in plaintext config/models.yaml by the included update_config script.
Persistence & Privilege
always:false and autonomous invocation defaults are normal. The notable persistence is that the skill will create and update a local config/models.yaml file containing API keys (unencrypted). The skill does not appear to modify other skills or system-wide settings, but storing secrets on disk and accepting them via chat increases exposure (chat logs, backups, shared environments).
What to consider before installing
This skill appears to be a legitimate multi‑model image/prompt tool, but it asks you to provide API keys and will write them unencrypted to pic-gen/config/models.yaml if you provide them by chat or via the set-key command. Before installing or using: (1) Prefer setting provider keys as environment variables rather than pasting them into chat; (2) Inspect and, if necessary, move the config file to a secure location and restrict permissions (e.g., chmod 600); (3) Do not commit config/models.yaml to any VCS or share it; (4) Be aware conversation logs might retain keys you paste into chat — if you did, rotate the exposed keys immediately; (5) Verify you trust the environment where the skill runs (shared or hosted agents may expose stored keys); (6) Note the registry metadata omission: confirm with the publisher why required env vars weren't declared. If you need stronger guarantees, run the scripts locally in an isolated environment and supply keys via environment variables rather than through the bot.Like a lobster shell, security has layers — review code before you run it.
latestvk9794k4awfabjwzxebcf5m49n183nxrs
License
MIT-0
Free to use, modify, and redistribute. No attribution required.