The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill is designed with security considerations for handling temporary credentials.
[COMMAND_EXECUTION]: The skill provides a PowerShell pattern for uploading files using Invoke-WebRequest. It includes specific instructions to suppress output using Out-Null to prevent sensitive upload URLs, headers, or tokens from being exposed in logs or terminal history.
[DATA_EXPOSURE]: The skill handles local image files and temporary upload secrets. It implements safe handling by advising against printing file contents or secrets in user-facing prose or saving them to project manifests.
[INDIRECT_PROMPT_INJECTION]: The skill ingests local file paths provided by users. While this is an ingestion surface for untrusted metadata, the skill restricts operations to metadata extraction and HTTP uploads, mitigating the risk of instructions embedded in file names or metadata being executed.

spritecook-upload-assets