The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (SAFE): The script scripts/juejin.js performs network requests to api.juejin.cn. While this is an external domain, it is the legitimate API for the service described in the skill. These operations are essential for the primary functionality of the skill.
[DATA_EXFILTRATION] (SAFE): No access to sensitive local files (like SSH keys, credentials, or environment variables) was detected. The data collected (article metadata) is intended for the user and is not sent to any suspicious third-party servers.
[COMMAND_EXECUTION] (SAFE): The script executes within the expected Node.js environment. No dangerous shell commands or privileged operations (like sudo) are used.
[REMOTE_CODE_EXECUTION] (SAFE): The skill does not download or execute remote scripts. It uses only built-in Node.js modules (https, zlib) and does not rely on external packages.
[PROMPT_INJECTION] (LOW): There is a potential for indirect prompt injection because the skill fetches untrusted article titles and summaries from a third-party website (Juejin). If an attacker publishes an article with malicious instructions in the title, the agent might attempt to follow them when processing the skill's output.
Ingestion points: scripts/juejin.js fetches data from api.juejin.cn via httpGet.
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided in the script output.
Capability inventory: The script has the capability to fetch and parse JSON data from the web.
Sanitization: No sanitization or filtering of the fetched article content is performed before outputting it to the agent.

juejin-article-trends