The Agent Skills Directory

[SAFE]: The skill consists entirely of architectural documentation and design guidelines. It does not include any executable scripts, binary files, or network-based functionality. No malicious patterns, such as credential theft or unauthorized command execution, were identified during the analysis.- [NO_CODE]: The skill is purely instructional and lacks any custom scripts (e.g., Python, JavaScript, Shell). This documentation-only approach minimizes the technical attack surface, as it relies on the agent's native reasoning rather than external code execution.- [INDIRECT_PROMPT_INJECTION]: The skill provides a framework for analyzing and refactoring code, which requires the agent to process user-provided source code. This represents an indirect injection surface where malicious instructions could be embedded in code comments or data. However, as the skill provides conceptual methodology rather than automated scripts for processing, this risk is inherent to the code-analysis task and is mitigated by the agent's default safety boundaries.
Ingestion points: Code modules and 'deepening candidates' provided by the user for analysis in DESIGN-IT-TWICE.md.
Boundary markers: The skill uses natural language instructions but does not specify technical delimiters for code processing.
Capability inventory: The skill uses the platform's standard Agent tool to spawn sub-tasks; it contains no internal scripts with file or network capabilities.
Sanitization: No specific sanitization or filtering logic is provided within the instructions.

codebase-design