The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's README.md includes a command for users to install the tool suite using 'curl | bash' from the author's GitHub repository. While common in developer workflows, this involves executing remote code. Additionally, the orchestrator executes local shell scripts defined in a manifest file to configure the project environment.
[EXTERNAL_DOWNLOADS]: The skill uses 'curl' to fetch version information and agent configuration files (such as .md and .mdc files) from the author's repository at 'raw.githubusercontent.com/Milkywayrules/verasic-skills'. These downloads are restricted to content from the skill's own vendor and are used to keep the local installation synchronized with upstream updates.
[COMMAND_EXECUTION]: The main script, 'init.sh', performs various shell operations including repository detection, integrity verification using SHA-256 hashes, and the invocation of sub-scripts for environment 'wiring' (e.g., setting git hooks and scaffolding environment files).
[DATA_EXFILTRATION]: The skill inspects the repository's git origin and file structure to determine the appropriate installation profile. It includes a security measure to strip potential credentials (tokens/passwords) from the remote origin URL before displaying it in generated reports, preventing sensitive data exposure in logs.
[SAFE]: All external resource references and network operations are directed to the author's own infrastructure. The skill includes explicit warnings and instructions regarding the safe management of GitHub Personal Access Tokens (PATs) and uses a confirmation-based workflow to prevent unauthorized changes.

verasic-init