The Agent Skills Directory

[SAFE]: The skill's primary function is to provide instructions and templates for Context Engineering (architectural documentation). It does not contain any executable scripts, network requests, or access to sensitive system resources.
[INDIRECT_PROMPT_INJECTION]: The skill includes functionality to ingest user decisions via the $ARGUMENTS variable and write them to markdown files (e.g., domain files) that are designed to be read by the agent in future sessions. This creates a surface for indirect prompt injection. However, the skill incorporates safety procedures such as requiring the agent to test the decision against the codebase and wait for owner approval before finalizing changes.

context-engineering