The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data retrieved from SAP HANA database objects, such as table names, column metadata, and query results.
Ingestion points: Data enters the agent context through commands like inspectTable, querySimple, and massConvert (described in references/command-reference.md and references/db-inspection.md).
Boundary markers: There are no explicit instructions or delimiters mentioned in the skill templates to prevent the agent from following instructions embedded within database content.
Capability inventory: The skill has the capability to execute system commands via hana-cli, write converted metadata files (ZIP/CDS), and interact with cloud APIs (documented in references/mass-operations.md and references/cloud-operations.md).
Sanitization: While references/development-environment.md describes SQL injection prevention functions like escapeDoubleQuotes, these are intended to protect the database from malicious input rather than protecting the agent from malicious data retrieved from the database.
[COMMAND_EXECUTION]: The skill facilitates the execution of numerous database management and inspection commands using the hana-cli tool, as detailed across the references/ directory.
[EXTERNAL_DOWNLOADS]: The skill documentation guides the user to download and install several external components.
Installs the hana-cli npm package from the public registry.
Executes install-btp.sh to fetch and install the SAP BTP CLI from an official SAP repository as described in references/development-environment.md.

sap-hana-cli