The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a shell script, scripts/audit-redux-store.sh, which performs a read-only scan of the project's source code using find and grep. The script implements security best practices, including set -euo pipefail and path traversal validation to prevent access outside the intended directory.
[EXTERNAL_DOWNLOADS]: The skill's instructions explicitly permit the use of standard package managers like npm for installing state management libraries while strictly prohibiting dangerous patterns such as curl | bash or unverified script execution.
[PROMPT_INJECTION]: The skill analyzes external application code, creating a surface for indirect prompt injection. This is addressed through a structured three-phase process (Inventory, Plan, Execute) that requires user confirmation and manual verification steps, reducing the risk of automated obedience to embedded instructions.
[DATA_EXFILTRATION]: The auditing process involves reading local source files to generate a migration report (redux-store-audit.md). All operations are performed locally, and the skill includes specific rules prohibiting the exposure of secrets or credentials in the generated reports.

redux-to-zustand-migrator