gemini-cli-security

Warn

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: MEDIUMOBFUSCATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [OBFUSCATION]: The skill's main analysis script uses hex-encoded character escapes within its regular expression patterns to represent sensitive keywords.\n
  • Evidence: In scripts/main.cjs, detection patterns for eval() use \x76 to represent the character 'v' (e.g., /e\x76al/).\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze arbitrary code files from the filesystem, which serves as a potential ingestion point for malicious instructions embedded in data.\n
  • Ingestion points: The scanFile function in scripts/main.cjs reads content from user-provided file paths during the scanning process.\n
  • Boundary markers: None. The script reads raw file content without applying delimiters or isolation instructions to the processing agent.\n
  • Capability inventory: The skill is configured with powerful tools including Bash, WebFetch, and Write.\n
  • Sanitization: No sanitization or filtering is performed on the ingested code content before analysis.\n- [EXTERNAL_DOWNLOADS]: The skill integrates with a well-known vulnerability database maintained by a trusted organization.\n
  • Evidence: The skill configuration and scripts/main.cjs reference the public OSV.dev API (https://api.osv.dev/v1/querybatch) for dependency scanning.\n- [COMMAND_EXECUTION]: The skill execution involves running shell commands and a Node.js script as part of its core operation.\n
  • Evidence: The SKILL.md file and usage examples describe running scripts/main.cjs via the node command.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 10, 2026, 12:13 PM