The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to execute various forms of the orca CLI tool (e.g., orca-dev, orca-ide, orca) based on environmental variables or system state.
[PRIVILEGE_ESCALATION]: The instructions explicitly tell the agent that if a command fails with runtime_access_denied, it should "re-run it with escalated permissions." This encourages the agent to bypass sandbox restrictions using mechanisms like sudo or administrator prompts.
[DYNAMIC_EXECUTION]: The skill uses a "discovery stub" pattern where the actual operational guide is loaded at runtime via the command ORCA skills get orca-cli. This means the agent's core instructions are fetched from an external executable rather than being statically defined in the skill.
[INDIRECT_PROMPT_INJECTION]: Because the skill fetches its instructions from the output of an external command (ORCA skills get orca-cli), it is vulnerable to indirect prompt injection if the executable or the data it retrieves is compromised or malicious.
Ingestion points: Output of the ORCA skills get orca-cli command.
Boundary markers: Absent; the skill does not specify delimiters or warnings to ignore instructions embedded in the command output.
Capability inventory: Shell command execution, environment variable resolution, and potential network access via the ORCA binary.
Sanitization: Absent; there is no validation or filtering mentioned for the content returned by the CLI before it is adopted as instructions.

orca-cli