astro-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust 'Safety Boundary' that explicitly prohibits file modifications (
Do not create, edit, move, or delete files), code execution (Do not run project code... tests, builds, checks, or scripts), and mutating Git operations. - [SAFE]: Data handling instructions treat external inputs (PR bodies, comments) as untrusted data and include explicit directives to ignore embedded instructions that would violate the review policy, effectively mitigating indirect prompt injection risks.
- [SAFE]: Network operations are restricted to read-only retrieval of pull request context using trusted tools like
ghandgitfor metadata only, with no evidence of data exfiltration or unauthorized external communication. - [SAFE]: The skill uses shell commands exclusively for read-only Git and GitHub operations (e.g.,
git diff,gh pr view). It prohibits shell pipelines or scripts for source code analysis to prevent command injection or execution of unverified logic. - [SAFE]: No obfuscation, persistence mechanisms, or privilege escalation attempts were detected. The skill is purely instructional and focused on static analysis tasks.
Audit Metadata