The Agent Skills Directory

[SAFE]: The skill implements a development philosophy centered on architectural documentation and testing. It includes various templates and guides that promote safe development practices.
[COMMAND_EXECUTION]: The skill utilizes local Python scripts (architecture_validator.py, check_coupling.py, and validate_risk_mitigation.py) to automate quality checks. These scripts perform static analysis on local files and do not exhibit any malicious behavior, such as unauthorized network access or sensitive file exposure.
[EXTERNAL_DOWNLOADS]: The provided Python scripts use the standard PyYAML library for parsing configuration files. This is a well-known and trusted dependency.
[PROMPT_INJECTION]: The skill is designed to process local documentation and source code for architectural review. Ingestion points: The static analysis scripts read local markdown and code files (e.g., SKILL.md, .py files). Boundary markers: No specific markers are used in the script outputs. Capability inventory: The scripts are limited to local file reading and logging to stdout. Sanitization: The AI interprets the text output of the scripts within the context of the workflow. Analysis: This functionality is inherent to the skill's primary purpose as a development tool and does not represent a malicious exploit attempt.

architect-first