The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to remain silent after certain actions, such as spawning tasks or completing checkpoints. These instructions are identified as UX optimizations designed to prevent chat noise, as the platform's user interface already displays the relevant progress cards and checklists to the user. This is a benign pattern rather than a concealment attempt.- [DATA_EXFILTRATION]: No unauthorized data access or network exfiltration patterns were identified. The skill manages context sharing between main and background agents using standard platform-provided variables and tags.- [COMMAND_EXECUTION]: The skill utilizes a set of recommended tools for workflow automation, including 'build-workflow' and 'executions'. These interactions are confined to the platform's intended toolset and do not involve arbitrary shell execution or unsafe command interpolation.

planned-task-runtime