The Agent Skills Directory

[SAFE]: The skill provides a structured framework for task decomposition and planning. All instructions and rules align with its stated purpose of improving delivery planning.
[EXTERNAL_DOWNLOADS]: No external network operations or remote resource fetches were identified in the instructions, scripts, or metadata.
[REMOTE_CODE_EXECUTION]: The provided shell scripts are used for local validation of the skill's output format. They do not execute remote code or perform unsafe operations.
[COMMAND_EXECUTION]: Shell script usage is limited to format validation and internal testing. Input handling within the scripts uses safe practices, such as quoted variables and heredocs, to prevent command injection.
[CREDENTIALS_UNSAFE]: No hardcoded credentials, API keys, or sensitive file paths were found. The skill does not include any secrets or instructions to handle them unsafely.
[PROMPT_INJECTION]: No attempts to override agent safety guidelines, extract system prompts, or bypass constraints were detected. The rules are focused on functional delivery logic and task management.
[DATA_EXFILTRATION]: There are no mechanisms for data exfiltration. The skill does not perform network requests or access sensitive user directories or environment variables.

delivery-engine