The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill accesses authentication cookies stored at the sensitive file path ~/.openclaw/secrets/xiaohongshu.json in check_comments.py, fetch_latest.py, and reply_fixed.py to authenticate browser sessions.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing untrusted user-generated comments from the Xiaohongshu platform.
Ingestion points: User comments are extracted from the page body via page.text_content('body') in check_comments.py and fetch_latest.py.
Boundary markers: No technical boundary markers or specific formatting are used to isolate untrusted data from the agent's instructions.
Capability inventory: The skill uses Playwright to perform automated browser actions such as clicking, typing, and sending messages, and can write debug information to local files.
Sanitization: Extracted comment content is processed and displayed without validation or sanitization.
[COMMAND_EXECUTION]: The skill utilizes Playwright to automate browser interactions and injects an external script (stealth.min.js) into the browser context to evade bot detection.

xiaohongshu-reply