The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified in the skill's instructions, metadata, or documentation.
[PROMPT_INJECTION]: No behavioral overrides, safety bypass attempts, or jailbreak patterns were detected. The skill instructions focus on defining a systematic research-first workflow for technical implementation.
[DATA_EXFILTRATION]: No unauthorized data access or harvesting patterns. All credentials and URLs used in code snippets and configuration examples are standard, non-sensitive placeholders (e.g., sk_live_xxx, api.example.com, DATABASE_URL).
[REMOTE_CODE_EXECUTION]: The skill references only official, well-known packages from the Svelte, SvelteKit, and Tailwind CSS ecosystems. No suspicious shell commands, piped execution (e.g., curl | bash), or dynamic loading from untrusted sources were found.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not utilize the !command syntax to execute shell commands at load time. Benign shell commands like find and npm are suggested for manual discovery and installation tasks by the agent.
[INDIRECT_PROMPT_INJECTION]: The skill defines a data ingestion surface by instructing the agent to read and synthesize its own documentation files. However, the content is static technical documentation provided by the author, and it includes best practices for security such as using DOMPurify for HTML sanitization and correctly scoping environment variables.

sveltekit-svelte5-tailwind-skill