The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data which could potentially contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Data enters the agent context through CLI flags such as --input, --content, and --notes in files like references/lark-okr-batch-create.md and references/lark-okr-create.md, which handle user-supplied objective titles and progress descriptions.
Boundary markers: The skill utilizes structured JSON formats (ContentBlock and SemiPlainContent) to handle rich text, which provides structural separation but lacks explicit instructions for the agent to ignore commands embedded within the text content.
Capability inventory: The agent can execute a wide range of commands via the lark-cli tool, including modifying OKR structures, managing progress records, and uploading local files via the +upload-image shortcut.
Sanitization: No explicit sanitization or filtering of user-provided text content is described before it is passed as arguments to the CLI tool.
[COMMAND_EXECUTION]: The skill relies on the lark-cli binary to perform all OKR-related tasks. While these operations are scoped to the okr subcommand and follow the vendor's official workflow, they involve executing shell commands with arguments derived from user-supplied input.

lark-okr