The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for setting up environment-specific tools and recommends several external packages for implementation tasks.
For image generation, it instructs the user to install the google-genai Python library via pip.
For CLI and TUI development, it provides a matrix of well-known libraries such as ink (Node.js), rich (Python), bubbletea (Go), and ratatui (Rust).
These recommendations target official registries (PyPI, NPM, Crates.io) and established open-source projects, which is consistent with the skill's primary purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requirements to generate prompts for external image-generation models (Google Gemini).
This creates an attack surface where a user could potentially supply instructions intended to bypass the image model's safety filters.
The skill mitigates this by including a dedicated image-policy recipe and reference/image-generation-content-safety.md, which defines a five-layer policy stack (pre-prompt filtering, post-generation classifiers, and persona refusals) to ensure compliance with content guidelines.
[COMMAND_EXECUTION]: The skill is designed to drive production-grade code implementation, including the generation of Python scripts, CLI tools, and build configurations.
It includes detailed guidance on authoring shell commands, handling exit codes, and cross-platform signal management.
The skill emphasizes safety practices like using timingSafeEqual for webhook verification and preventing ReDoS (Regular Expression Denial of Service) by auditing regex patterns against untrusted input.

builder