The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a self-contained Python script to interact with CSDN APIs via standard system calls. Access is scoped to the user's provided session cookies.
[DATA_EXPOSURE]: The skill documentation explicitly warns against printing or echoing the CSDN_COOKIES environment variable. The script implements careful cookie handling, ensuring credentials are only sent to the legitimate csdn.net and csdnimg.cn domains.
[REMOTE_CODE_EXECUTION]: While the script can download external images to re-host them on CSDN's CDN, it includes a robust SSRF guard. It validates that image URLs are public, resolves hostnames to verify they are not private/loopback/link-local addresses, and blocks redirects to prevent bypassing these checks.
[PROMPT_INJECTION]: The skill implements a gated execution model for publishing articles. The AI agent is instructed to always perform a dry-run first and requires an explicit user confirmation before appending the --confirm flag to execute write operations.

csdn