The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting untrusted data from the codebase being analyzed.\n
Ingestion points: The skill uses an exploration tool to walk the codebase and reads various files like CONTEXT.md, ADRs, and source code files.\n
Boundary markers: There are no explicit instructions or delimiters used to separate the analyzed code content from the agent's core instructions.\n
Capability inventory: The skill has the ability to write to the local filesystem (temporary directory) and execute shell commands (git, open, xdg-open, start).\n
Sanitization: The skill does not describe any sanitization steps for the ingested code content. Furthermore, the generated HTML report configures the Mermaid diagramming tool with securityLevel: "loose", which allows the rendering of HTML and scripts within diagrams, potentially enabling cross-site scripting (XSS) if malicious strings are present in the codebase comments or file names.\n- [COMMAND_EXECUTION]: The skill executes several system commands to perform its analysis and present findings.\n
It runs git log --oneline to analyze repository history.\n
It uses platform-specific commands such as open (macOS), xdg-open (Linux), and start (Windows) to display the generated HTML report.\n- [EXTERNAL_DOWNLOADS]: The generated report includes external dependencies from well-known content delivery networks.\n
It loads the Tailwind CSS styling framework from cdn.tailwindcss.com.\n
It loads the Mermaid diagramming library from cdn.jsdelivr.net.\n
These downloads are used for the primary purpose of rendering the visual report.

vibe-refactor