exa-reference-architecture

Pass

Audited by Gen Agent Trust Hub on Mar 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a Retrieval-Augmented Generation (RAG) pattern that is inherently susceptible to indirect prompt injection.
  • Ingestion points: Untrusted data from external web sources is ingested via the exa.searchAndContents tool and processed in the ragSearch function in SKILL.md.
  • Boundary markers: The implementation uses basic delimiters (---) and source headers ([Source X]), which provide minimal protection against instructions embedded within the retrieved content.
  • Capability inventory: No dangerous operations (such as filesystem writes or subprocess execution) are present in the provided script examples.
  • Sanitization: There is no evidence of sanitization, filtering, or validation of the external content before it is interpolated into the prompt context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 24, 2026, 04:38 PM