The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads and searches content from external spreadsheets via sheets +read and sheets +find, which can introduce untrusted data into the agent's context.
Ingestion points: Data retrieved from spreadsheet cells using lark-cli sheets +read and lark-cli sheets +find (documented in references/lark-sheets-read.md and references/lark-sheets-find.md).
Boundary markers: Absent. The instructions do not specify any delimiters or safety warnings for the agent when processing retrieved cell content.
Capability inventory: The agent has capabilities to execute shell commands (lark-cli), write to the local filesystem (sheets +export --output-path), and modify permissions on cloud documents.
Sanitization: Absent. The skill does not provide methods for sanitizing or validating spreadsheet data before it is processed.
[PRIVILEGE_ESCALATION]: The skill contains instructions in references/lark-sheets-create.md for the agent to grant full_access (administrator) permissions to users for newly created spreadsheets. Although the instructions emphasize user confirmation and reporting, this involves the management of high-privilege access rights.
[COMMAND_EXECUTION]: The skill relies on the execution of a local binary lark-cli to perform all operations, which is a standard pattern for this type of integration but involves direct shell interaction.

lark-sheets