agentcore-websearch

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the agentcore-websearch CLI tool from a trusted GitHub repository hosted by the aws-samples organization (github.com/aws-samples/sample-agentcore-websearch-agent-skill).
  • [COMMAND_EXECUTION]: The skill executes local shell commands to verify credentials (aws sts get-caller-identity), check the environment (command -v), and perform search operations (agentcore-websearch).
  • [PROMPT_INJECTION]: The skill processes untrusted data from web search results, presenting a surface for indirect prompt injection. 1. Ingestion points: Results returned by the agentcore-websearch command in SKILL.md. 2. Boundary markers: The skill does not provide specific instructions to the agent to treat search results as untrusted or to ignore embedded instructions. 3. Capability inventory: The agent can execute shell commands for searching and diagnostics. 4. Sanitization: No evidence of sanitization or filtering of search results is defined in the skill's instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 02:00 AM
Security Audit — agent-trust-hub — agentcore-websearch