The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill explicitly instructs the agent to re-run commands with escalated permissions if a runtime_access_denied error occurs. This encourages the agent to bypass sandbox restrictions using mechanisms like sudo or administrative shells.
[COMMAND_EXECUTION]: The skill requires the execution of various CLI binaries (orca, orca-dev, orca-ide) based on environment variables. It also executes skills get which dynamically retrieves further instructions or documentation into the session.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it ingests untrusted data from an external source.
Ingestion points: The command ORCA skills get orca-emulator fetches a version-matched guide that is then processed by the agent.
Boundary markers: None specified in the instructions to distinguish between the fetched guide and system instructions.
Capability inventory: The skill allows arbitrary shell command execution via the resolved ORCA executable.
Sanitization: No sanitization or validation of the fetched guide content is performed before processing.

orca-emulator