The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the official aliyun CLI and dataworks-public plugin to manage cloud resources.
Evidence: Instructions include standard operations such as aliyun dataworks-public create-semantic-job and aliyun plugin install.
[EXTERNAL_DOWNLOADS]: The skill downloads semantic analysis artifacts from Alibaba Cloud's infrastructure.
Evidence: The scripts/download_results.py script manages downloads from *.aliyuncs.com domains.
Security Controls: The script enforces HTTPS, validates that the hostname belongs to Alibaba Cloud, checks for path traversal in the output directory, and ensures files are written with restricted permissions (umask 077).
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from user-supplied resource identifiers and API responses.
Ingestion Points: Resource names, project IDs, and API response payloads in SKILL.md and scripts/download_results.py.
Boundary Markers: Instructions mandate the use of JSON encoders for shell arguments and dedicated scripts for handling API output, avoiding direct interpolation.
Capability Inventory: CLI command execution and file system writes to a user-specified directory.
Sanitization: The scripts/download_results.py script provides extensive sanitization, including control character removal from filenames and strict path resolution logic.

alibabacloud-dataworks-semantic