The Agent Skills Directory

[PROMPT_INJECTION]: The skill's evaluation configuration (references/evals.yaml) contains prompt injection test strings, such as instructions to 'Ignore previous instructions' and 'include raw credentials'. While these are intended for safety validation, they are detected as injection patterns.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including git diffs, pull request descriptions, and markdown implementation plans. It lacks explicit boundary delimiters or 'ignore' instructions for this content, creating a surface for indirect prompt injection where a malicious document could influence agent behavior.
[EXTERNAL_DOWNLOADS]: The HTML templates download well-known visualization and animation libraries (Mermaid.js, Chart.js, and Anime.js) from established CDN services. These dependencies are standard for the skill's purpose of generating standalone browser-based visualizations.
[COMMAND_EXECUTION]: The skill workflow involves executing various developer tools via the command line, including git, gh (GitHub CLI), grep, and wc. It also uses mise and uv to manage Python environments for supporting scripts.

visual-explainer