The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions are focused on helping the agent understand user requirements for prototyping. There are no patterns of behavior override, safety filter bypass, or jailbreak attempts detected.
[DATA_EXFILTRATION]: The skill explicitly advises against persistence and suggests using local scratch databases or in-memory state for prototypes. There are no network calls to unknown external domains or attempts to access sensitive system files (e.g., SSH keys, AWS credentials).
[REMOTE_CODE_EXECUTION]: The instructions involve generating standard HTML, JavaScript, and TypeScript code for local use by the developer. It does not perform any remote script downloads or piped command executions from untrusted sources.
[COMMAND_EXECUTION]: While the skill mentions running prototypes via standard task managers (pnpm, python, bun), these are described as examples of project workflows rather than arbitrary command execution. No suspicious or privileged commands (like sudo) are present.
[OBFUSCATION]: The content is presented in clear, plain-text markdown. No Base64, zero-width characters, homoglyphs, or other obfuscation techniques were identified.
[DYNAMIC_EXECUTION]: The skill provides templates for UI switchers and logic modules. These are static code generation tasks intended for developer review and do not involve unsafe deserialization or runtime code injection from external inputs.

prototype