together-embeddings

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified in the RAG pipeline pattern.
  • Ingestion points: External document data is ingested via the documents parameter in scripts/embed_and_rerank.py and scripts/embed_and_rerank.ts.
  • Boundary markers: Absent. The example in SKILL.md interpolates context directly into the system prompt without delimiters or instructions to the AI to ignore instructions within that context.
  • Capability inventory: The skill performs network requests to Together AI's official API endpoints.
  • Sanitization: Absent. No filtering or escaping is performed on the retrieved content before it is processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 08:06 PM