vn-tone
Pass
Audited by Gen Agent Trust Hub on May 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from users and interpolates it into sub-agent prompts without sufficient isolation.
- Ingestion points: User-provided Vietnamese text or file paths enter the agent context as the
briefvariable in theSKILL.mdorchestrator. - Boundary markers: Absent. The orchestration logic does not specify the use of delimiters (e.g., XML tags, triple backticks) or 'ignore embedded instructions' warnings for the untrusted input passed to the Diagnostic, Polisher, and Critic agents.
- Capability inventory: The skill environment has high-privilege tools enabled, including
Bash,WebFetch,Read,Grep, andGlob, which could be targets for manipulation via injected instructions. - Sanitization: Absent. There is no evidence of validation or filtering performed on the input text to strip potential instruction-like patterns before processing.
Audit Metadata