The Agent Skills Directory

[SAFE]: The skill acts as a comprehensive developer's toolkit for creating and validating ChatGPT plugins. It emphasizes security best practices, such as credential management and privacy policy compliance, and includes specific logic to detect hardcoded secrets in plugin manifests.\n- [COMMAND_EXECUTION]: The provided Python scripts (check_package.py and test_check_package.py) are designed for local validation of plugin assets. They utilize standard library modules and optional PyYAML for safe data parsing. The scripts do not perform network requests, spawn unauthorized subprocesses, or execute arbitrary shell commands.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a data ingestion surface by reading user-provided plugin configuration files (JSON, YAML, Markdown). It mitigates potential injection risks by using safe parsing methods (json.loads, yaml.safe_load) and restricting its operations to static validation rather than execution of the content.

chatgpt-plugin