The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the aliyun CLI and the lhm plugin from official Alibaba Cloud domains (aliyuncli.alicdn.com). It also installs the uv package manager from its official source (astral.sh). These operations are necessary for environment preparation and use trusted or well-known service domains.
[COMMAND_EXECUTION]: The skill frequently executes shell commands using subprocess.run() to interact with the aliyun CLI and various database command-line tools (such as curl for Presto). This is the core functionality required for performing SQL DryRun validations, data consistency checks, and workflow migrations.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it ingests untrusted data from user-supplied SQL files and configuration files without explicit boundary markers or sanitization, while maintaining powerful capabilities like command execution and network access.
Ingestion points: Reads user SQL files for conversion and validation (e.g., in references/lhm-sql-conversion/scripts/dryrun.py).
Boundary markers: Absent. The scripts do not use specific delimiters to isolate the content of processed SQL files.
Capability inventory: Executes shell commands via aliyun CLI, performs database network operations, and writes files to the local system.
Sanitization: Absent. The scripts perform dialect-specific text replacements but do not filter for natural language instructions embedded within the SQL or configuration data.

alibabacloud-lakehouse-migration