skills/notque/vexjoy-agent/translate/Gen Agent Trust Hub

translate

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection because it ingests untrusted text from user documents and processes it through parallel sub-agent calls.
  • Ingestion points: Untrusted source text is ingested from the user's document as described in Phase 1 and Phase 3 of SKILL.md.
  • Boundary markers: The sub-agent dispatch pattern in references/modes.md uses a triple-dash delimiter to separate instructions from content, but it does not include an explicit safety instruction to the LLM to ignore any commands or instructions found within the text chunk.
  • Capability inventory: The skill is capable of writing translation results to the local filesystem (SKILL.md Phase 3) and dispatching sub-task calls to other LLM instances (references/modes.md).
  • Sanitization: There is no evidence of filtering, escaping, or validation of the input text before it is interpolated into the sub-agent prompt.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 09:19 PM
Security Audit — agent-trust-hub — translate