skills/drolu/agent-skills/orca-cli/Gen Agent Trust Hub

orca-cli

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a local orca CLI to manage terminals and automate browser sessions, including starting terminals with arbitrary commands and sending text input to them.
  • [COMMAND_EXECUTION]: Provides capabilities to execute JavaScript within a browser page context (orca eval) and run extended browser commands (orca exec), which allows for flexible environment control.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it reads and processes potentially untrusted data from terminal outputs and browser accessibility snapshots.
  • Ingestion points: terminal output (orca terminal read) and browser snapshots (orca snapshot).
  • Boundary markers: None.
  • Capability inventory: Terminal command execution (orca terminal send), browser interaction (orca click), and JavaScript evaluation (orca eval).
  • Sanitization: None.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 07:11 PM
Security Audit — agent-trust-hub — orca-cli