lark-whiteboard

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs extracting and then inserting a "画板 Token" directly into CLI commands (e.g., --whiteboard-token ), which requires the model to handle and potentially output secret token values verbatim, an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md requires fetching user-provided document URLs with lark-cli (e.g., "lark-cli docs +fetch --doc --as user") and parsing returned content/tokens and Dry Run output (checking for "XX whiteboard nodes will be deleted") as part of the mandatory upload workflow, so it clearly ingests user-generated third‑party content that can materially change subsequent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs remote code at runtime via npx/npm (e.g., "npx -y @larksuite/whiteboard-cli@^0.1.0" and "npm install -g @larksuite/whiteboard-cli@^0.1.0"), which downloads and executes an external package that the skill declares as a required dependency, so it meets the criteria for a runtime-executed external dependency.