code-review
Fail
Audited by Snyk on Nov 18, 2026
Risk Level: CRITICAL
Full Analysis
CRITICAL E004: Prompt injection detected in skill instructions.
- Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate CodeRabbit review workflow, but its Antigravity sandbox handoff has the user run a command that sets
proceed-in-sandboxand broader sandbox permissions, which weakens a protection beyond what the review task needs, though it is disclosed and not malicious.
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (medium risk: 0.30). The optional
coderabbit pullrequest <number-or-url> --show-prompts --agentworkflow, which the skill description advertises as "GitHub PR fix prompts," feeds CodeRabbit-generated prompts derived from GitHub PR review text that outside contributors can post to.
MEDIUM W013: Attempt to modify system services in skill instructions.
- Attempt to modify system services in skill instructions detected (low risk: 0.30). The skill is a code-review workflow that mostly stays within read-only review actions, but it allows a retry outside the sandbox on the host after a failed sandbox attempt, which is a mild, ambiguous bypass of an execution boundary.
Issues (3)
E004
CRITICALPrompt injection detected in skill instructions.
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
W013
MEDIUMAttempt to modify system services in skill instructions.
Audit Metadata