skills/gek0v/orca/orca-cli/Gen Agent Trust Hub

orca-cli

Fail

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: HIGHPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The instructions explicitly direct the agent to attempt to bypass security sandbox restrictions by escalating its own permissions.
  • Evidence: The text in SKILL.md states, "If it fails with runtime_access_denied, your sandbox blocked the connection: re-run it with escalated permissions."
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to dynamically ingest instructions from an external command-line tool at runtime, which creates an attack surface for malicious instructions to be injected into the agent's context.
  • Ingestion points: The agent is instructed to run ORCA skills get orca-cli to "load the version-matched guide" which then guides the agent's subsequent actions.
  • Boundary markers: None. There are no instructions to the agent to validate the loaded content or to ignore any embedded commands within the tool's output.
  • Capability inventory: The agent is granted broad capabilities via the orca CLI, including operating worktrees, terminals, repositories, and an embedded browser.
  • Sanitization: None. The skill lacks any mechanisms to filter or sanitize the content loaded from the CLI output before it is processed by the AI.
  • [COMMAND_EXECUTION]: The skill facilitates the execution of local shell commands via the orca executable, including logic for resolving the executable path across different environments (WSL, dev, Linux).
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Oct 7, 2026, 10:46 PM