skills/kerenski/orca/orca-cli/Gen Agent Trust Hub

orca-cli

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions the agent to execute a local command-line interface (CLI) named orca (or variants like orca-ide, orca-dev, orca-cli).
  • The agent resolves the executable path based on environment variables or the operating system environment.
  • Subcommands are used to manage local worktrees, terminals, and an embedded browser instance.
  • The skill includes safety guidance to prevent accidental execution of the GNOME Orca screen reader on Linux systems.
  • [PROMPT_INJECTION]: The skill implements a dynamic context loading pattern by instructing the agent to retrieve its full usage guide from the output of a local command (ORCA skills get orca-cli).
  • Ingestion points: Instructions are ingested into the agent's context from the stdout of a local binary.
  • Boundary markers: No explicit boundary markers or 'ignore' instructions are used for this dynamic content; the agent is told to treat the output as the authoritative guide.
  • Capability inventory: The skill has the capability to execute shell commands and interact with the local file system through the Orca toolset.
  • Sanitization: There is no sanitization of the guide content as it is expected to be provided by the verified local toolchain.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 06:13 AM
Security Audit — agent-trust-hub — orca-cli