The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data from Lark documents and user comments while maintaining high-impact capabilities such as file deletion and permission modification, creating a surface for indirect prompt injection attacks.
Ingestion points: The skill retrieves external content through comment listings (lark-drive-reactions.md) and document fetching mechanisms (referenced in lark-drive-add-comment.md).
Boundary markers: Absent. The instructions do not include markers or directives to separate retrieved data from agent instructions or to ignore embedded commands.
Capability inventory: The skill possesses extensive write and delete capabilities, including +delete (file/folder removal), +move (file relocation), and permission.members (permission grants and owner transfers).
Sanitization: Absent. There is no evidence of data validation or sanitization before external content is processed by the agent.
[COMMAND_EXECUTION]: The skill utilizes the lark-cli binary to perform all Drive operations. This includes administrative actions like transfer_owner and permission.members.create, which are consistent with the skill's primary purpose of cloud storage management.

lark-drive