The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables management of cloud infrastructure by executing shell commands via the tcb CLI, covering deployment, authentication, and resource configuration across all reference files.\n- [DYNAMIC_EXECUTION]: The skill facilitates the execution of dynamic SQL and NoSQL (MongoDB-style) commands against cloud databases. It includes safety mitigations such as recommending --read-only modes for queries and requiring user confirmation before executing data mutations, as detailed in references/mysql.md and references/nosql.md.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by ingesting and processing data from external sources such as cloud logs and database records.\n
Ingestion points: Log search operations (tcb logs search, tcb fn log) and database queries (tcb db execute, tcb db nosql execute) found in references/functions.md, references/mysql.md, and references/nosql.md.\n
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the ingested data.\n
Capability inventory: The skill has extensive shell execution capabilities, including resource deployment, file system management, and database modification.\n
Sanitization: There is no evidence of sanitization or filtering logic applied to external data before it is presented to the agent context.

cloudbase-cli