The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses rg (ripgrep) to perform read-only scans of the project's codebase. These commands are limited to identifying installed dependencies in package.json and searching for specific technical keywords in source files (e.g., 'three', 'gsap', 'canvas'). This is a standard reconnaissance pattern for repository-focused skills and does not involve executing untrusted code.
[EXTERNAL_DOWNLOADS]: The skill references official documentation and technical blogs for Three.js, GSAP, and React Three Fiber. These external resources are standard industry references and do not involve downloading executable scripts or binary payloads during skill operation.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it ingests untrusted data from the user's repository via code searches and possesses the capability to modify project files.
Ingestion points: The skill reads file contents and dependency lists using rg in SKILL.md (Phase 1: Recon).
Boundary markers: No specific delimiters or "ignore instructions" warnings are applied to the data retrieved from the codebase.
Capability inventory: The skill is designed to implement features by writing or modifying source code in the project.
Sanitization: The skill does not explicitly describe sanitization or validation of the content found within the files before using it to inform implementation decisions.

enhance-web-web3d