The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses platform-specific commands (xdg-open on Linux, open on macOS, and start on Windows) to automatically open a generated HTML report in the user's default browser.
[EXTERNAL_DOWNLOADS]: The generated HTML report fetches the Tailwind CSS framework from cdn.tailwindcss.com and the Mermaid.js diagramming library from cdn.jsdelivr.net. These are well-known services used for the skill's primary presentation purpose.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection as it processes untrusted data from a codebase to generate visual reports.
Ingestion points: The skill reads project files, CONTEXT.md, and Architecture Decision Records (ADRs) from the target codebase.
Boundary markers: There are no explicit instructions to the agent to sanitize or delimit content extracted from the codebase before placing it into the HTML template.
Capability inventory: The skill has the capability to write files to the temporary directory and execute system commands to open those files.
Sanitization: No specific sanitization or escaping of codebase content is mentioned. Notably, the Mermaid configuration in HTML-REPORT.md sets securityLevel: "loose", which allows the execution of scripts or tags embedded within the generated diagram data, potentially enabling XSS if the codebase contains malicious diagram definitions.

improve-codebase-architecture