finance-tax-helper
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No evidence of prompt injection, jailbreak attempts, or safety filter bypasses was found in the instructions. The content is consistently focused on tax guidance.
- [DATA_EXFILTRATION]: No unauthorized data collection or external transmission patterns were identified. The skill incorporates a mandatory step (Step 0) to mask Korean personal identifiable information (PII) before processing user input, which reduces the risk of sensitive data exposure.
- [OBFUSCATION]: No hidden text, encoded commands, homoglyph attacks, or concealed URLs were detected in the skill instructions or reference files.
- [REMOTE_CODE_EXECUTION]: The skill does not download or execute external code. It utilizes static local markdown files and refers to verified peer skills within the same modu-ai vendor ecosystem.
- [INDIRECT_PROMPT_INJECTION]: Although the skill processes user-provided financial data, it implements a structured sanitization workflow involving PII masking and a post-processing review chain to mitigate potential injection risks.
Audit Metadata