The Agent Skills Directory

[Security-Aware Instructions]: The skill explicitly instructs the agent to treat external inputs (Git diffs, PR descriptions) as untrusted data and warns against following instructions found within them. This is a proactive measure against indirect prompt injection.
[Sensitive Data Verification]: The instructions include a specific lens for checking sensitive data exposure in logging and error messages, which aligns with security best practices for code review.
[Trusted Resource Reference]: The skill references the official Google Go Style Guide (google.github.io), which is a well-known and trusted technical resource.
[Standard Tooling]: The skill utilizes standard development tools such as git and lsp for code analysis, which is typical for a code review workflow.

adk-go-self-review