The Agent Skills Directory

[SAFE]: The skill consists of several Python scripts that perform static analysis on local source code to generate architecture diagrams and dependency reports. These scripts use standard library modules and do not perform network operations, shell execution, or sensitive data access beyond the directory provided by the user. The code is well-structured and follows diagnostic best practices, including limits on file processing to prevent resource exhaustion.
[INDIRECT_PROMPT_INJECTION]: The skill reads untrusted content from local files during the analysis process. While this represents a potential surface for indirect prompt injection (e.g., via malicious instructions hidden in code comments), the risk is minimal as the scripts only extract specific metadata via regex and do not execute or blindly display the entire content.
Ingestion points: scripts/architecture_diagram_generator.py, scripts/dependency_analyzer.py, and scripts/project_architect.py use Path.read_text() to process files in the target directory.
Capability inventory: The scripts are limited to file system reads and report writing; no network or subprocess capabilities are present.
Boundary markers: Output is formatted as structured reports or code blocks within the agent's response.
Sanitization: Content is filtered through regular expressions to extract specific architectural metadata (imports, class names, etc.), effectively stripping out arbitrary instructional text.

senior-architect