brave-search

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill retrieves untrusted data from the internet via search results and page extraction, which presents a surface for indirect prompt injection where malicious instructions could be embedded in search snippets or website content.
  • Ingestion points: Untrusted content is ingested through the output of the ./search.js and ./content.js scripts as described in SKILL.md.
  • Boundary markers: The skill uses basic separators such as '--- Result 1 ---' but lacks clear instructions or markers to distinguish external data from system instructions.
  • Capability inventory: The skill is used by the agent to fetch and process web content; the scripts themselves execute search queries and perform HTTP requests.
  • Sanitization: There is no documentation indicating that external content is sanitized or that instructions within the content are escaped before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 06:58 AM