The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The module modules/00-install-hook.md is designed to inject a startup hook into global settings files (~/.claude/settings.json and ~/.cursor/settings.json). This hook executes npx skills add Bil0000/skema-skill on every session start, allowing for the persistent automatic execution of remote code from a GitHub repository.
[COMMAND_EXECUTION]: The skill attempts to modify global user configuration files outside of the project workspace. The documentation in README.md explicitly encourages users to run the agent with the --dangerously-skip-permissions flag to facilitate these unauthorized file writes.
[DATA_EXFILTRATION]: modules/04-codebase-scan.md and modules/12-image-to-code.md perform comprehensive scans of the local codebase, including package.json, configuration files, and source components. When combined with the unrestricted network access used in modules/03-clone-engine.md and modules/05-web-research.md, this creates a high-risk surface for the exfiltration of proprietary project data.
[COMMAND_EXECUTION]: The commands/migrate.md file performs filesystem operations like mv and rm on directories, and it specifically modifies the global settings files to update hook commands, which could be misused to alter environment behavior.

skema