glossary-term-translations
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project strings and translations through the list_strings and list_language_translations tools, creating a potential surface for indirect prompt injection. This risk is effectively mitigated by clear boundary markers instructing the agent to treat data as non-executable. The skill's primary capability is the add_term tool, which is sanitized via a human-in-the-loop presentation of results and a restriction to draft status, preventing unauthorized or automated modifications to the glossary.
- [SAFE]: No malicious patterns or security risks were identified. The workflow adheres to security best practices, including the principle of least privilege for write operations and maintaining a strict data-instruction boundary.
Audit Metadata