The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled Python script (scripts/fetch_ai_lab.py) to scrape and process project data. It also provides a fallback instruction using curl for manual data fetching.
[EXTERNAL_DOWNLOADS]: Fetches project metadata and data snapshots from Alibaba Cloud's official domains (aliyun.com and aliyuncs.com). As these originate from a well-known service and the skill author, they are classified as safe vendor resources.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes data from an external web page to populate the agent's response.
Ingestion points: Data is retrieved from www.aliyun.com via the fetch_ai_lab.py script.
Boundary markers: The skill uses a strictly ordered template (①-⑤) to structure the output, which helps differentiate between instructions and data.
Capability inventory: The agent can execute local Python scripts, read/write to its memory tool, and create scheduled tasks (cron jobs) on various supported platforms.
Sanitization: The fetching script performs basic HTML unescaping and text normalization, but does not provide specific protection against adversarial instructions embedded in the remote project descriptions.
[SAFE]: The Python script fetch_ai_lab.py includes an SSL certificate verification bypass as a fallback mechanism (ssl._create_unverified_context). While this is a security best-practice violation, it is noted here as a minor concern given the trusted nature of the target domains.

alibabacloud-ai-innovation-lab-skill