invoice-tax-checklist
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of natural language instructions and YAML metadata. It does not include any Python scripts, Node.js packages, shell commands, or other executable files, which eliminates direct code execution risks.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external data, including transaction descriptions and invoice images, to generate checklists. This represents a potential indirect prompt injection surface. However, the instructions mitigate this by requiring a strict separation between user-provided information and agent confirmations, and by providing a rigid structure for the generated Word document. The instructions also explicitly prevent the agent from interpreting the authenticity of images or providing conclusive tax opinions, further limiting the impact of potentially malicious input data.
Audit Metadata