The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a command-line interface for database management. It includes extensive safety features such as a mandatory --write flag for any data-modifying operations and a block on UPDATE/DELETE statements that lack a WHERE clause.
[CREDENTIALS_UNSAFE]: While the skill manages authentication for database and OSS access (e.g., hologres volume create), it does so through standard profile-based configuration stored locally in ~/.hologres/config.json. There are no hardcoded secrets or evidence of credential harvesting.
[DATA_EXFILTRATION]: The skill facilitates data movement between Hologres and Alibaba Cloud OSS (Object Storage Service). These operations are part of the core functionality for data import/export and are directed at the user's own cloud infrastructure. No unauthorized data transmission to third-party domains was detected.
[DYNAMIC_EXECUTION]: The test suite (tests/test_skill_completeness.py) utilizes importlib.import_module() to dynamically inspect the CLI codebase. This is used solely to verify that the documentation remains consistent with the actual source code and does not represent a runtime risk for the agent.
[EXTERNAL_DOWNLOADS]: The skill references the hologres-cli package, which is hosted on a vendor-owned registry. These references are consistent with standard installation procedures for developer tools.

hologres-cli