The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references and provides instructions for installing a full bundle from the developer's official domain at https://genjutsu.athevon.dev.
[COMMAND_EXECUTION]: The skill executes various shell commands (ls, grep, find) to perform automated discovery of the project's technical stack, including identifying web frameworks, Android (Compose), and Apple (SwiftUI) environments.
[COMMAND_EXECUTION]: The skill invokes local Python scripts (search.py and audit.py) to perform design system lookups and project-wide design audits.
[DYNAMIC_EXECUTION]: The skill implements a dynamic loading system (load_skill) that resolves and reads markdown modules from multiple local filesystem paths (e.g., /mnt/skills, ~/.claude/skills, /sessions) to provide context-specific design guidance.
[INDIRECT_PROMPT_INJECTION]: The skill scans untrusted project files to inform its configuration, presenting a potential attack surface.
Ingestion points: Reads content from package.json, build.gradle, Package.swift, and other project metadata files.
Boundary markers: None explicitly implemented for ingested file content.
Capability inventory: Possesses Bash, Read, Write, and Edit tools, enabling it to modify the project based on the analyzed data.
Sanitization: No specific sanitization techniques are detailed for the ingested project configuration data.

paint