The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill explicitly instructs the agent to "re-run it with escalated permissions" if a runtime_access_denied error occurs. This encourages the agent to bypass security sandbox constraints or use administrative privileges (e.g., sudo).
[COMMAND_EXECUTION]: The skill directs the agent to execute system-level binaries (orca, orca-ide, or orca-dev) based on environment variables. This pattern can be dangerous if the environment paths or variables are not strictly controlled, potentially leading to the execution of unintended or malicious code.
[DYNAMIC_EXECUTION]: The skill implements a "discovery stub" pattern where it instructs the agent to run ORCA skills get orca-cli to fetch and load additional instructions at runtime. This allows for the dynamic injection of content that was not present in the static SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to instructions embedded in external tool output. Ingestion points: Output of the ORCA skills get orca-cli command described in SKILL.md. Boundary markers: Absent; the agent is told to treat the output as a "version-matched guide" without constraints. Capability inventory: Terminal management, worktree operations, repository interaction, and browser control. Sanitization: Absent; there are no instructions to validate or filter the content returned by the tool.

orca-cli