The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill instructs the agent to re-run commands with escalated permissions if a 'runtime_access_denied' error is encountered, which could lead to unauthorized administrative actions on the host system.
[DYNAMIC_EXECUTION]: The skill uses a dynamic instruction loading pattern where it executes ORCA skills get orca-emulator to fetch and load a version-matched guide at runtime, potentially injecting new instructions into the agent's context.
[COMMAND_EXECUTION]: The skill resolves and executes local binaries (orca, orca-ide, orca-dev) based on environment variables like ORCA_CLI_COMMAND and ORCA_DEV_REPO_ROOT, which could be manipulated to execute unintended files if the environment is compromised.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect injection by processing output from an external command as instructions.
Ingestion points: The output of the ORCA skills get orca-emulator command (SKILL.md).
Boundary markers: Absent; there are no specified delimiters to separate the loaded guide from the agent's primary instructions.
Capability inventory: Execution of shell commands via the resolved binary and simulator control tools (simctl, xcodebuild).
Sanitization: None mentioned; the agent is instructed to follow the guide provided by the command output.

orca-emulator