SiliconFlow Image Gen

Generate images using SiliconFlow API (FLUX.1, Stable Diffusion, etc.)

This skill appears to do what it says: send prompts to SiliconFlow and save resulting images. Before installing, verify the skill source (there are minor metadata mismatches and the top-level registry said no required env var while the files do require SILICONFLOW_API_KEY). Confirm you trust api.siliconflow.cn and are willing to provide that API key. Ensure curl is available on the host (the script calls curl via subprocess). Inspect or run the included scripts/generate.py yourself; check permissions on ~/.openclaw/openclaw.json (it may contain other keys) and avoid placing unrelated high-privilege secrets there. If you want stricter isolation, consider running the script manually rather than granting an agent autonomous invocation.

Type: OpenClaw Skill Name: siliconflow-image-gen Version: 1.0.0 The OpenClaw AgentSkills skill bundle is classified as benign. The `scripts/generate.py` script correctly implements its stated purpose of generating images via the SiliconFlow API. It accesses the `SILICONFLOW_API_KEY` from environment variables or `~/.openclaw/openclaw.json` as declared in `SKILL.md` and `skill.json`. Network calls to `api.siliconflow.cn` and subsequent image downloads are performed using `subprocess.run` with a list of arguments, effectively preventing shell injection vulnerabilities from user-controlled inputs (prompt, model, size, output path). No evidence of intentional harmful behavior, data exfiltration beyond the stated API interaction, persistence mechanisms, or prompt injection attempts against the agent were found in any of the files.