query

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The bin/query script downloads a platform-specific binary and its release information from https://agent-cdn.nutrient.io. This is a vendor-owned resource for pspdfkit-labs (Nutrient) used to provide the skill's core search functionality.
  • [COMMAND_EXECUTION]: The skill executes several local shell commands to manage the lifecycle of the search binary, including mkdir for directory creation, tar for archive extraction, ln -s for symlink management, and exec to launch the search tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from user-provided Markdown or text files (INPUT.md), which presents a vulnerability surface where malicious instructions embedded in those documents could influence the agent's behavior when search results are returned.
  • Ingestion points: The skill reads data from the file path provided in the INPUT.md argument.
  • Boundary markers: While the output includes line number references (e.g., Lines A-B), there are no explicit security delimiters or warnings used to isolate the searched content from the agent's instructions.
  • Capability inventory: The skill executes a local binary to perform search operations; it does not grant the input data direct access to network or file-write capabilities.
  • Sanitization: No sanitization or filtering is performed on the text windows retrieved from the file before they are presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 11:27 PM
Security Audit — agent-trust-hub — query