The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository, including source code, commit history, and documentation, which could influence agent behavior. * Ingestion points: Scans codebase files, commit logs via git log, and reads GLOSSARY.md and Architecture Decision Records (ADRs) in docs/adr/. * Boundary markers: The skill does not implement boundary markers or explicit instructions to ignore instructions embedded within the codebase data. * Capability inventory: Writing HTML reports to the temporary directory, modifying repository files (GLOSSARY.md), and executing shell commands for repository inspection and file opening. * Sanitization: No sanitization of ingested codebase content is performed before it is used to generate reports or guide sub-agent exploration.- [COMMAND_EXECUTION]: The skill invokes shell commands to analyze the project environment and present reports to the user. * Evidence: Executes git log --oneline to identify codebase hotspots and uses system commands (xdg-open, open, or start) to launch the generated HTML report.
[EXTERNAL_DOWNLOADS]: The generated HTML report incorporates styling and diagramming libraries from established third-party services. * Evidence: Fetches Tailwind CSS from cdn.tailwindcss.com and Mermaid.js from cdn.jsdelivr.net. These are well-known services used for visual representation of the architectural review.

improve-codebase-architecture