The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill instructs the agent to bypass environment restrictions. Specifically, it states that if an operation fails with a 'runtime_access_denied' error due to a sandbox blocking the connection, the agent should 're-run it with escalated permissions'. This encourages the agent to seek higher privileges (such as sudo) to circumvent security boundaries.
[COMMAND_EXECUTION]: The skill revolves around executing various local binaries (orca, orca-dev, orca-ide, orca-cli). It uses these tools for terminal control, worktree management, and running arbitrary shell commands.
[DYNAMIC_EXECUTION]: The skill uses a 'discovery stub' pattern where it avoids including its own usage guide. Instead, it instructs the agent to run ORCA skills get orchestration to fetch its instructions dynamically at runtime. This means the actual operational logic of the skill is determined by the output of a local binary rather than the reviewed markdown file.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection.
Ingestion points: The agent ingests instructions from the output of the ORCA skills get orchestration command.
Boundary markers: The instructions do not define delimiters or warnings to treat the fetched content as untrusted data.
Capability inventory: The skill has the capability to run shell commands, manage terminal sessions, and control worktrees via the resolved ORCA executable.
Sanitization: There is no mention of escaping or validating the dynamic instructions returned by the CLI before the agent follows them.

orchestration