The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches frontend libraries from well-known and trusted services, including jsDelivr, ObservableHQ, and Google Fonts. These resources (Mermaid.js, Chart.js, D3.js, Observable Plot, and various fonts) are used solely to render visualizations within the generated HTML files. These references align with the skill's primary purpose and follow established safety standards for web resources.
[COMMAND_EXECUTION]: The skill executes several common command-line utilities to perform its analysis. This includes git for retrieving diffs and logs, gh for interacting with GitHub Pull Requests, wc and grep for data extraction, mmdc for Mermaid diagram validation, and open/xdg-open for displaying the final results. It also optionally utilizes surf-cli for generating AI-based conceptual illustrations. These operations are limited to the agent's runtime environment and are necessary for the skill's documented features.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill processes untrusted data from git diff outputs, repository source code, and user-provided implementation plans.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for this external content.
Capability inventory: The agent has capabilities to run shell commands (git, gh, surf, mmdc), write to the local file system (saving HTML and design briefs), and open files in a web browser.
Sanitization: There is no explicit sanitization step described for the ingested text before it is analyzed or rendered.
Assessment: While an attack surface exists, the risk is assessed as safe given the skill's intended use-case of helping developers visualize their own work and the limited scope of the agent's actions.
[DYNAMIC_EXECUTION]: The skill generates HTML files that include inline JavaScript logic for interactive UI components such as diagram zooming, panning, and progressive disclosure panels. This code is transparent and focused on enhancing the user experience of the visual diagrams.

visual-explainer