parallel-web-search

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute an installation script directly from the vendor's official domain using curl -fsSL https://parallel.ai/install.sh | bash. This is a standard setup procedure for the provided toolset.
  • [EXTERNAL_DOWNLOADS]: Installs the parallel-web-tools package from public registries using pipx. This dependency is consistent with the vendor's naming conventions and the skill's stated purpose.
  • [COMMAND_EXECUTION]: Utilizes parallel-cli to perform search queries, retrieve JSON results, and save them to the local filesystem for processing. This is the primary function of the skill.
  • [DATA_EXFILTRATION]: Instructs the user on how to set a PARALLEL_API_KEY for authentication. While this handles a credential, it is used exclusively for legitimate communication with the vendor's search service.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it fetches and processes untrusted data from the web.
  • Ingestion points: Web search results fetched by parallel-cli in SKILL.md.
  • Boundary markers: None; the skill directly summarizes content from search excerpts without delimiters or instructions to ignore embedded commands.
  • Capability inventory: Subprocess execution via parallel-cli (Bash) in SKILL.md.
  • Sanitization: None; results are parsed and synthesized directly into the response.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 01:01 AM