The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions focus on project structure and documentation generation. There are no attempts to bypass safety filters, extract system prompts, or use adversarial role-play patterns.
[EXTERNAL_DOWNLOADS]: The skill does not perform any remote downloads. All utility scripts and templates are part of the skill's own package and are copied locally.
[DATA_EXFILTRATION]: No network operations or credential-harvesting behaviors were identified. The skill only interacts with the local file system to organize project assets.
[COMMAND_EXECUTION]: The skill includes utility scripts intended for local execution. These scripts (read-yaml.mjs, validate-yaml.mjs, etc.) use standard Node.js modules like 'fs' and 'path' to perform transparent tasks without executing arbitrary shell commands or remote code.
[SAFE]: The skill follows standard development workflows for project bootstrapping, including tech stack detection, directory creation, and generating .gitignore files. All provided scripts are well-documented and functional as described.

cc-init