The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform various command-line operations to manage the development lifecycle.
Uses git add and git commit for version control within the plan implementation tasks.
Uses pytest for automated testing and verification of the generated code steps.
Utilizes environment-specific tools like orca and dispatch for creating and injecting tasks into an orchestration system (e.g., orca orchestration task-create, dispatch --inject).
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its design as a data propagator.
Ingestion points: Reads specification data from files located at llmwiki/wiki/sources/draft/DDMMYY-<tên>.md.
Boundary markers: No explicit boundary markers or instruction-isolation delimiters are defined for the imported content.
Capability inventory: The skill has the ability to write to the filesystem, commit to Git, and dispatch automated tasks to headless agents via specific CLI tools.
Sanitization: Content from the source specification is copied verbatim into the implementation plan, which is then used to drive subsequent agent actions. The instructions specifically forbid summarization, which facilitates the propagation of content from the source into the final task briefs.

plan