tui-button
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill is composed entirely of Markdown instructions, JSON schemas, and usage examples. It does not include any scripts, binaries, or automated installation steps that would execute on the host system.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external data through its 'Input Model' JSON schema. While this could theoretically ingest malicious instructions embedded in component properties, the risk is negligible as the skill lacks access to sensitive tools, file systems, or network operations.
- Ingestion points: The
Input Modelsection inSKILL.mddefines how user-provided JSON properties are mapped to the UI. - Boundary markers: The skill uses strict
### OUTPUTheaders to separate rendered content, but lacks explicit 'ignore instructions' warnings for the input data. - Capability inventory: No tools or dangerous capabilities (such as
subprocess,fetch, orfile_write) are requested or used. - Sanitization: No specific sanitization or escaping logic is defined for the input properties beyond ASCII truncation rules.
Audit Metadata