The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to interact with the environment, specifically running git log --oneline to identify frequently changed files and platform-specific commands like xdg-open, open, or start to display the generated HTML report to the user. These operations are performed within the local development context.
[EXTERNAL_DOWNLOADS]: The generated HTML report references third-party libraries including Tailwind CSS and Mermaid.js from well-known content delivery networks (Tailwind CDN and JSDelivr). These resources are used exclusively for styling and rendering architectural diagrams within the generated report.
[PROMPT_INJECTION]: The skill is exposed to indirect prompt injection via the codebase content it processes.
Ingestion points: The skill reads source code files, domain models (CONTEXT.md), architecture decision records (ADRs), and git commit history.
Boundary markers: No specific delimiters are defined to separate codebase content from agent instructions during the analysis phase.
Capability inventory: The skill can spawn sub-agents for codebase exploration, write HTML files to the system temporary directory, and execute system commands to open files.
Sanitization: The skill does not explicitly sanitize data extracted from the codebase before inclusion in the HTML report. The Mermaid.js initialization uses securityLevel: "loose", which allows rendering of HTML tags and presents a potential local cross-site scripting (XSS) surface if malicious code comments are included in the generated diagrams.

improve-codebase-architecture