tui-chat-group
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted JSON data from an input model and interpolates it into structured output blocks, including drawing commands for a downstream tool.
- Ingestion points: Untrusted content enters the skill's context through the
props,modelValue, andstate.errorfields of the Input Model defined in SKILL.md. - Boundary markers: The skill employs distinct markdown headers (e.g.,
### OUTPUT: TUI_RENDER) and triple-backtick code blocks to separate user data from instructions, which provides some structural delimitation. - Capability inventory: The skill generates specifications and batch design calls intended for a
Pencil MCPtool, creating a capability chain where generated output influences another tool's actions. - Sanitization: The instructions do not define any sanitization, escaping, or validation logic for the input fields before they are rendered into the potentially executable
PENCIL_BATCH_DESIGNblock.
Audit Metadata