code-search

Warn

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the 'bunx' utility which, by default, downloads the 'codebasesearch' package from the npm registry if it is not already present in the environment.- [REMOTE_CODE_EXECUTION]: The use of 'bunx' results in the execution of remotely-sourced code from a public package registry at runtime.- [COMMAND_EXECUTION]: The skill explicitly allows the execution of bash commands through the 'Bash(bunx codebasesearch*)' tool configuration.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it retrieves and processes content from the codebase, which is considered untrusted external data.
  • Ingestion points: Files within the codebase searched by the 'codebasesearch' utility.
  • Boundary markers: No delimiters or instructions are provided to the agent to distinguish between codebase content and system instructions.
  • Capability inventory: The skill is permitted to execute bash commands.
  • Sanitization: There is no logic to sanitize or filter potential instructions embedded within the codebase files before they are read by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 5, 2026, 12:07 AM