The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the lark command-line utility to list tables, fields, and records from Lark Bitable apps. This requires the binary to be present in the environment's PATH.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and processes untrusted data from external Bitable records.
Ingestion points: Untrusted data enters the agent's context through the lark bitable records command output in SKILL.md.
Boundary markers: While the output is structured as JSON, which provides a logical separation, it does not prevent the agent from being influenced by instructions embedded within the field values of retrieved records.
Capability inventory: The agent has the capability to execute shell commands via the lark tool based on the data retrieved.
Sanitization: No explicit sanitization, filtering, or escaping of the retrieved database content is documented or implemented before the data is processed by the agent.

bitable