create-judge
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates a workflow where the agent ingests and reviews untrusted data from LLM spans, creating an attack surface for indirect prompt injection.\n
- Ingestion points: The agent is instructed to review evaluation results and reasons via the dashboard or API (Step 7 in SKILL.md).\n
- Boundary markers: There are no instructions for the agent to use delimiters or ignore embedded instructions when reading evaluation data.\n
- Capability inventory: The agent can modify platform configuration by creating, updating, and linking judges via REST API endpoints (references/defaults-and-api-reference.md).\n
- Sanitization: The instructions do not include steps for sanitizing the evaluated span content or validating the integrity of the evaluation feedback.
Audit Metadata