The Agent Skills Directory

[SAFE]: The skill implements a per-slice code review workflow that promotes security best practices, such as specifying 'security-sensitive' review modes that trigger audits for common vulnerabilities (OWASP Top 10, credential leakage, input validation).
[SAFE]: The scripts/review-package.sh script is a legitimate utility for generating git patches for review. It uses standard git commands and restricts output to local directory structures within the .omo/ folder.
[SAFE]: The usage of sub-agents ('oracle' or 'unspecified-high') for review is a design choice to ensure independent verification, which is a common pattern in AI agent orchestration.
[SAFE]: The skill uses allowed-tools: "Read Bash Grep", which are standard, low-privilege tools required for reading project files and executing the provided shell scripts.

slice-review