The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the Alibaba Cloud CLI installation script and binary packages from the official vendor CDN (aliyuncli.alicdn.com). These resources originate from a well-known service provider.
[REMOTE_CODE_EXECUTION]: The installation guide provides instructions to execute the vendor's official CLI installation script via a piped bash command. This operation is directed at a trusted vendor source.
[COMMAND_EXECUTION]: The script scripts/pds_poll_processor.py invokes the aliyun CLI using subprocess.run with a list of arguments and shell=False. This implementation follows security best practices to prevent command injection.
[EXTERNAL_DOWNLOADS]: Utility scripts download analysis results and associated images from signed URLs pointing to Alibaba Cloud storage infrastructure (aliyuncs.com).
[PROMPT_INJECTION]: The skill processes untrusted data from external files, creating an indirect prompt injection attack surface.
Ingestion points: Document and video analysis results (summaries, keywords, transcripts) are downloaded and processed by scripts/doc_analysis_formatter.py and scripts/video_analysis_formatter.py.
Boundary markers: The scripts do not implement explicit boundary markers or instructions to ignore embedded prompts within the processed content.
Capability inventory: The skill possesses capabilities for network operations (via the requests library), shell command execution (via the aliyun CLI in scripts/pds_poll_processor.py), and local file system access.
Sanitization: While content is structured via JSON parsing, the textual content is not subjected to specific sanitization or filtering before being presented to the agent.

alibabacloud-pds-multimodal-search