SiliconFlow Video Gen

Generate videos using SiliconFlow API with Wan2.2 model. Supports both Text-to-Video and Image-to-Video.

This skill appears to do what it says: it needs a SiliconFlow API key and will send prompts (and optional image URLs) to api.siliconflow.cn to request video generation. Before installing or providing an API key: 1) Verify the SiliconFlow API domain (api.siliconflow.cn) is the legitimate service you expect. 2) Prefer to create a restricted API key (scoped or rate-limited) rather than using a high-privilege key. 3) Review scripts/generate.py yourself (it’s short and readable) to confirm there are no additional network destinations. 4) Correct the metadata mismatch (registry-level 'required env' vs skill.json) or ask the publisher to fix it — mismatched metadata is usually a bookkeeping error but should be clarified. 5) If you need stricter isolation, run the script in a sandbox or environment with only the minimal network and file access necessary.

Type: OpenClaw Skill Name: siliconflow-video-gen Version: 1.0.1 The skill bundle is designed to generate videos using the SiliconFlow API. It reads the API key from expected locations (`SILICONFLOW_API_KEY` environment variable or `~/.openclaw/openclaw.json`) and makes a network request to the hardcoded `https://api.siliconflow.cn/v1/video/submit` endpoint using `curl`. The `prompt` and `image_url` inputs are JSON-encoded before being passed to `curl`, mitigating direct shell injection risks. There is no evidence of data exfiltration to unauthorized endpoints, persistence mechanisms, obfuscation, or malicious prompt injection attempts in `SKILL.md` or `README.md`. The behavior aligns with the stated purpose.