The Agent Skills Directory

[DATA_EXPOSURE]: The skill documentation describes searching for and reading environment configuration files such as .env to retrieve database credentials. This is standard behavior for a configuration generation tool.
Evidence: SKILL.md contains discovery commands like fd -t f ".env*" . to locate secret files.
[COMMAND_EXECUTION]: The skill utilizes shell commands for environment discovery and testing database connectivity.
Evidence: SKILL.md references the use of clickhouse-client and fd for pre-implementation discovery.
[INDIRECT_PROMPT_INJECTION]: The skill processes configuration data from the local environment and file system, which represents a surface for indirect prompt injection if the source data is attacker-controlled.
Evidence Ingestion points: scripts/generate_dbeaver_config.py (reads from environment variables).
Evidence Boundary markers: None explicitly defined for interpolated environment values.
Evidence Capability inventory: The skill has the capability to write files (scripts/generate_dbeaver_config.py) and execute bash commands.
Evidence Sanitization: The skill uses Pydantic v2 for type validation and schema enforcement in scripts/generate_dbeaver_config.py and scripts/validate_config.py.

clickhouse-pydantic-config