The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes bash commands using the tai CLI tool (e.g., tai tool task_list). While this is the intended primary function, passing parameters like board_id or assistant_id directly into a shell command string creates a potential surface for command injection if user-supplied input is not strictly validated or escaped before execution.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes data from Kanban boards and tasks. This data ingestion surface is vulnerable to indirect prompt injection; if task descriptions or board names contain malicious instructions, the agent could potentially be manipulated when it processes the tool's output. The skill lacks explicit boundary markers or instructions to ignore embedded commands within the fetched data.
Ingestion points: Data enters the agent context via the board_list and task_list tool outputs (SKILL.md).
Boundary markers: None present; the skill does not instruct the agent to distinguish between tool metadata and task content.
Capability inventory: The skill can execute bash commands and read board/task data (SKILL.md).
Sanitization: No evidence of sanitization or escaping of the task data before it is returned to the agent.

yao-board