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 is encountered. This encourages the agent to acquire higher system privileges to bypass security restrictions.
[DYNAMIC_EXECUTION]: The instructions direct the agent to load a 'version-matched guide' by executing ORCA skills get orca-per-workspace-env. This pattern causes the agent to retrieve and execute new instructions from an external source at runtime, which are not present in the skill's static files.
[COMMAND_EXECUTION]: The skill facilitates the execution of local CLI tools (orca, orca-ide, orca-dev) based on the session environment. While these are vendor-provided tools, they are used to perform the dynamic instruction loading and environmental setup.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface where it ingests untrusted data from an external command output to use as further instructions. * Ingestion points: The output of the ORCA skills get command is used as a guide for subsequent session steps. * Boundary markers: The skill does not define specific delimiters or instructions to ignore potential commands embedded in the fetched guide. * Capability inventory: The agent is granted the capability to execute shell commands via the ORCA executable. * Sanitization: There is no evidence of sanitization or validation performed on the content fetched from the CLI tool before the agent processes it as instructions.

orca-per-workspace-env