The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents a vulnerability surface for indirect prompt injection because its core functionality includes reviewing and auditing external code supplied by users.
Ingestion points: The agent ingests untrusted data in the form of source code files and diffs when operating in review, audit, or improve modes.
Boundary markers: There are no specific instructions in the agent's prompts to use delimiters or security warnings when processing this untrusted code.
Capability inventory: The agent has access to powerful tools including bash, write_to_file, and view_file across all supported platforms.
Sanitization: The skill lacks explicit sanitization or filtering protocols for the content it reads from the file system or user input before analysis.
[COMMAND_EXECUTION]: The skill is configured to use the bash tool to perform technical tasks and provides documentation for command-line operations.
Evidence: references/threejs.md suggests using npx @gltf-transform/cli for asset optimization. While this is a standard developer workflow, the availability of shell access in a skill that processes external code requires user oversight.
[EXTERNAL_DOWNLOADS]: The documentation identifies and suggests the installation of several third-party animation and 3D libraries.
Evidence: The skill references external dependencies such as gsap, lenis, animejs, three.js, and matter-js. These are well-known and widely used libraries within the web development community.

animate