The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script (scripts/juejin.py) executed via Bash to interact with the Juejin API. The script is self-contained and uses only the Python standard library.
[CREDENTIALS_SAFE]: The skill handles authentication using JUEJIN_COOKIES, which are injected as an environment variable by the platform's connector. The instructions and code explicitly warn against printing or echoing these secrets. The request function uses add_unredirected_header to prevent the cookie from being leaked to third-party domains during HTTP redirects.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. The skill operates locally on the user's account using urllib for API calls.
[EXTERNAL_DOWNLOADS]: All network operations are directed to the official api.juejin.cn endpoint. No untrusted external dependencies are downloaded or executed.
[PROMPT_INJECTION]: The skill instructions include safety gates for write operations. The publish command requires an explicit --confirm flag as the final argument, which prevents accidental or injected publishing actions without user verification.
[DATA_EXFILTRATION]: Data access is limited to the user's own Juejin profile and articles. There are no patterns suggesting the exfiltration of sensitive local data (like SSH keys or AWS credentials) to external servers.

juejin