ensemble-solving
Fail
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: HIGHPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection (Category 8) due to the way it handles user-provided tasks.
- Ingestion points: In
SKILL.md(Step 2), the user's original task and context are passed directly to theensemble-orchestratortool. - Boundary markers: There are no defined delimiters or instructions to prevent the subagents from obeying commands embedded within the user's task.
- Capability inventory: The skill is intended for high-impact tasks like code generation, refactoring, and architecture decisions, which could lead to the generation of malicious code or insecure designs if the subagents are compromised.
- Sanitization: No sanitization or validation logic is present to filter malicious instructions from the user input before processing.
Recommendations
- AI detected serious security threats
Audit Metadata