The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation recommends executing various shell commands and OCI CLI operations for environment configuration and troubleshooting. Examples include oci session authenticate, oci vault vault list, and chmod 600 ~/.oci/oci_api_key.pem to secure private key files.
[EXTERNAL_DOWNLOADS]: The skill instructions advise users to install the official oci Python package via pip and run the oracle.oci-vault-mcp-server using the uvx tool runner to enable MCP capabilities.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it retrieves secret data from external OCI Vault storage which could potentially contain malicious instructions embedded within secret values.
Ingestion points: Secret values and metadata are ingested into the agent context via tools such as get_secret_value and get_secret as described in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the retrieved secret content.
Capability inventory: The skill provides comprehensive secret management capabilities, including the ability to read, create, update, and delete secrets.
Sanitization: There are no explicit sanitization or filtering steps described for the data retrieved from the vault service.

oci-vault