The Agent Skills Directory

[COMMAND_EXECUTION]: Spawns background worker processes using subprocess.Popen in scripts/cli/worker_utils.py and scripts/cli/dual_logger.py to handle asynchronous data analysis tasks without blocking the main agent turn. This is a legitimate implementation of the skill's primary concurrency model.
[COMMAND_EXECUTION]: Invokes external notification tools such as openclaw or clawdbot via subprocess.run in scripts/cli/notify.py to push status updates to the user session.
[EXTERNAL_DOWNLOADS]: Interacts with official Alibaba Cloud API and OSS endpoints (aliyuncs.com) for session management, data analysis, and report file transfers in scripts/data_agent/client.py and scripts/data_agent/file_manager.py. These represent vendor-owned resources consistent with the skill's purpose.
[PROMPT_INJECTION]: Processing of external databases and user-uploaded files (CSV, Excel) in scripts/cli/cmd_file.py and scripts/cli/cmd_db.py creates a surface for indirect prompt injection where data content could attempt to influence the analysis engine.
Ingestion points: Ingests untrusted data through the file and db subcommands via the scripts/cli/cmd_file.py and scripts/cli/cmd_db.py modules.
Boundary markers: Relies on delimiters provided by the display formatters and backend service-level isolation rather than explicit local boundary markers.
Capability inventory: Possesses capabilities for local command execution (workers/notifications) and broad network access to Aliyun services.
Sanitization: Content is transmitted to the analysis engine without local sanitization, relying on the safety guardrails of the underlying Alibaba Cloud Data Agent service.

alibabacloud-data-agent-skill