choose-your-parallel-api

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for an agent to perform search and research tasks using live web data from external sources, which constitutes a significant attack surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters the agent context via the /v1/search and /v1/extract API endpoints described in the skill, which return search excerpts and full-page content from third-party websites.
  • Boundary markers: The skill provides no instructions on using boundary markers, XML tags, or explicit 'ignore' instructions to separate retrieved web content from the agent's core instructions.
  • Capability inventory: The skill encourages the agent to 'Implement the smallest working path' and 'verify one representative input', which typically involves the use of file-editing and command-execution tools that could be abused if the agent follows malicious instructions from a webpage.
  • Sanitization: There is no mention of sanitizing, escaping, or pre-processing the retrieved web content to remove potential injection vectors before the agent reasons over the data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 08:07 PM
Security Audit — agent-trust-hub — choose-your-parallel-api