parallel-web-search

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a command-line interface tool (parallel-cli) to perform web searches and manage service balances.
  • Evidence: Found in SKILL.md using the Bash tool to run parallel-cli search and parallel-cli balance commands.
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external content from the web, which creates a surface for potential indirect instructions embedded in search results.
  • Ingestion points: External web content is fetched via parallel-cli and saved to temporary JSON files in /tmp/ before being read by the agent.
  • Boundary markers: The instructions do not define specific delimiters or "ignore instructions" tags for the retrieved search result excerpts.
  • Capability inventory: The agent has access to the Bash tool and can write to the local file system (/tmp/).
  • Sanitization: No explicit sanitization or filtering of the retrieved web content is specified before the agent synthesizes its response.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 10:25 PM
Security Audit — agent-trust-hub — parallel-web-search