oracle

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill metadata specifies the installation of the @steipete/oracle package from the NPM registry.
  • [COMMAND_EXECUTION]: The documentation provides instructions for running the oracle CLI and npx commands to bundle files, manage sessions, and interact with AI models.
  • [COMMAND_EXECUTION]: Includes instructions for starting a remote server using oracle serve, which opens a network listener on the host machine.
  • [DATA_EXFILTRATION]: The tool's core functionality is to transmit local file content to external AI services. The documentation explicitly advises users to exclude sensitive files like .env and authentication tokens.
  • [PROMPT_INJECTION]: This skill presents an indirect prompt injection surface as it ingests untrusted local files via the --file parameter (File: SKILL.md). No boundary markers or sanitization methods are documented to mitigate instructions embedded in processed files. The skill has capabilities to perform network operations and execute CLI commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:15 AM