judgment-day

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill performs 'Pattern 0: Skill Resolution' which ingests rules from '.ai/atl/skill-registry.md' and other 'SKILL.md' files within the repository. This distilled content is interpolated into the prompts for judge and fix sub-agents, creating a surface where malicious instructions in project files could potentially influence agent behavior.
  • Ingestion points: Pattern 0 reads local configuration and documentation files (SKILL.md, .ai/atl/skill-registry.md) from the target project.
  • Boundary markers: Distilled rules are placed in a 'Project Standards' block but lack robust delimiters or safety instructions to prevent the sub-agents from obeying embedded commands.
  • Capability inventory: The skill utilizes sub-agent orchestration (via the 'task' or 'Task' tools) to perform code analysis and apply file-system modifications.
  • Sanitization: There is no evidence of sanitization or strict schema validation for the data gathered from the project files before it is injected into the sub-agent contexts.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 11:46 AM
Security Audit — agent-trust-hub — judgment-day