The Agent Skills Directory

[SAFE]: The skill employs secure orchestration patterns, including delegated task isolation and defined iteration limits to prevent uncontrolled execution loops.
[PROMPT_INJECTION]: An indirect prompt injection surface exists as the skill processes external code files through sub-agent prompts. 1. Ingestion points: Target files passed to Judge and Fix sub-agents. 2. Boundary markers: Absent in the delegation templates. 3. Capability inventory: The delegated Fix Agent has the capability to modify source code. 4. Sanitization: No sanitization is performed on the code content before processing. This risk is inherent to the functional purpose of automated code review and does not represent malicious intent.

judgment-day