The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute the standard terminal command date at the start of the workflow. This is used to replace placeholders in the generated document with the current system time, which is a legitimate and safe operation for documentation tasks.
[SAFE_PRACTICE]: The instructions implement strict operational constraints, such as requiring the agent to read templates fresh from disk and waiting for user confirmation after each step, which reduces the risk of hallucinations or state drift.
[DATA_INGESTION]: The skill processes user-provided information (Epic titles, descriptions, business value) to populate a Markdown template. While no explicit sanitization is mentioned, the data is strictly used for document generation rather than being passed to executable contexts, minimizing injection risks.

012-agile-epic