web-search

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize untrusted data from the internet via a search tool, which presents a standard attack surface for indirect prompt injection.
  • Ingestion points: Data enters the context through the output of the WebSearch tool as defined in the allowed-tools section of SKILL.md.
  • Boundary markers: The instructions do not specify explicit delimiters (like XML tags or triple quotes) to separate search results from the system instructions.
  • Capability inventory: The skill's capabilities are limited to WebSearch. It lacks higher-risk permissions such as file-system writes, network exfiltration tools (curl/wget), or subprocess execution.
  • Sanitization: No explicit sanitization or filtering of external content is described in the workflow.
  • [SAFE]: The skill follows least-privilege principles by restricting tool access to WebSearch only. No evidence of credential exposure, obfuscation, or remote code execution was found.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:49 PM
Security Audit — agent-trust-hub — web-search