The Agent Skills Directory

[PROMPT_INJECTION]: No evidence of instructions attempting to bypass safety filters or override agent behavior was found. The language is purely technical and instructional.
[DATA_EXFILTRATION]: No sensitive file paths, credentials, or network exfiltration patterns were detected. All code examples operate within the local application context (Three.js scene).
[REMOTE_CODE_EXECUTION]: No patterns of remote script execution or dynamic code evaluation (eval/exec) were found. The skill relies on well-known, versioned libraries.
[EXTERNAL_DOWNLOADS]: The skill mentions official packages from @thatopen, which is consistent with the skill's purpose and the author's identity. No suspicious downloads from untrusted sources were detected.
[COMMAND_EXECUTION]: No dangerous shell commands or privilege escalation attempts (sudo/chmod) were identified.
[DYNAMIC_EXECUTION]: The skill does not generate or execute code at runtime from untrusted inputs. Examples focus on standard library API calls.
[OBFUSCATION]: No obfuscated strings, hidden characters, or encoded URLs were found across the analyzed files.

thatopen-impl-highlighting