Use Jimeng AI 4.0 (Volcengine) to generate images from text or image references, and optionally send results to Feishu.

This skill appears to be a legitimate wrapper for Volcengine (Jimeng) image generation, but there are two issues you should consider before installing or running it: (1) the scripts require VOLCENGINE_AK and VOLCENGINE_SK (secret credentials) even though the registry metadata did not declare them — verify you trust the skill source before providing keys and prefer an account/key with limited scope and short lifetime; (2) when asked to deliver results to Feishu the script posts a JSON payload to http://localhost:18789/message (a local listener), not directly to Feishu's official API. If you don't already run a trusted local bridge that forwards to Feishu, that localhost endpoint could capture or forward generated image URLs and captions unexpectedly. Recommendations: inspect the scripts (you already have them), confirm the skill author/source, run the skill in an isolated environment (or container) if you must supply real credentials, consider creating an ephemeral/limited Volcengine key, and verify or replace the localhost callback with a delivery mechanism you control. Additional information that would raise confidence: a known homepage or repository, registry metadata corrected to declare required env vars, and an explanation of the expected local Feishu bridge or an option to post directly to Feishu with explicit credentials.

Type: OpenClaw Skill Name: jimeng-t2i-i2i-image-generator Version: 0.1.0 The skill exhibits a client-side JSON injection vulnerability in `scripts/jimeng.sh`. User-provided inputs such as `prompt`, `reference_url`, and `target` are directly embedded into JSON payloads for `curl` requests to `visual.volcengineapi.com` and `http://localhost:18789/message` without proper sanitization. This could lead to malformed JSON requests, API errors, or potentially unexpected behavior if the receiving APIs are not robust against such input. While there is no clear evidence of intentional malicious behavior like data exfiltration or persistence, this input sanitization flaw is a significant vulnerability.