The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the oci Command Line Interface to manage cloud resources, such as inference endpoints, notebook sessions, and model deployments. This behavior is entirely consistent with the skill's stated purpose of providing documentation for OCI AI services.
[EXTERNAL_DOWNLOADS]: The skill references the installation of Python packages (oci, openai, oci-ai-speech-realtime) and Conda environments (generalml_p38_cpu_v1) required for interacting with OCI. These are documented as standard setup steps within the OCI ecosystem.
[PROMPT_INJECTION]: The skill describes workflows that ingest data from external sources like OCI Object Storage for processing by AI models (e.g., document extraction, speech transcription). This constitutes an indirect prompt injection surface typical for AI-integrated applications.
Ingestion points: OCI Object Storage buckets (e.g., images, docs, audio) used as input for vision and speech jobs.
Boundary markers: No specific delimiters or markers are mentioned in the command-line examples; the skill relies on service-level parsing.
Capability inventory: The skill utilizes the shell tool for OCI CLI operations, which can be used to read or modify cloud resources.
Sanitization: No explicit sanitization or filtering logic is provided in the instructions, as the data is handled by managed cloud services.

oci-ai