judgment-day
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a coordinator that ingests untrusted data from the project environment and user input, then passes it to sub-agents (Judges and Fix Agents) via prompt interpolation.
- Ingestion points: Files and features specified as the 'target' in SKILL.md, the contents of the
.atl/skill-registry.mdfile, and custom user-provided review criteria. - Boundary markers: The sub-agent prompt templates for the Judge and Fix Agent do not utilize explicit delimiters or 'ignore' instructions to prevent the sub-agents from being influenced by malicious instructions embedded within the target files or registry rules.
- Capability inventory: The orchestrator launches sub-agents using the
delegatetool, and the Fix Agent is specifically designed to perform file modifications based on instructions derived from the review process. - Sanitization: The skill lacks logic to sanitize or validate the external content before it is interpolated into the
{describe target}and{paste matching compact rules blocks}placeholders.
Audit Metadata