Pretext Three.js frontend skill

Create dynamic web experiences with physical-text layouts using Pretext for precise text measurement and Three.js for interactive 3D rendering and animation.

This pack appears coherent and instruction-only (no installs or credential requests). Before using: 1) Note provenance: the skill's source/homepage is unknown—prefer to inspect generated code and verify any third-party packages (e.g., @chenglou/pretext, three) before installing them. 2) When an agent uses these prompts to produce runnable projects, review package.json/pinned dependency versions and avoid executing npm/yarn installs from untrusted registries without review. 3) Check any generated code for unexpected network calls or endpoints before running in a production environment; run prototypes in a sandbox. 4) If you plan to have the agent publish artifacts or run CI, confirm CI secrets are not exposed to generated scripts. Overall: the skill itself is coherent and low-risk, but vet any code the agent produces and the dependencies it installs.

Type: OpenClaw Skill Name: pretext-agents-skill Version: 1.0.0 The skill bundle is a legitimate set of instructions and code patterns for building interactive, typography-focused frontend experiences using Three.js and the Pretext layout engine. The content, including SKILL.md and EXAMPLES.md, focuses on creative UI/UX design, performance optimization, and specific architectural patterns (e.g., Textura, Breaker). There are no indicators of malicious intent, data exfiltration, or harmful prompt injection.