The Agent Skills Directory

[CREDENTIALS_UNSAFE]: Provides dedicated scripts (get-token.sh and get-token.ps1) to securely retrieve IMA credentials from platform-managed environments or local configuration files in standard directories (~/.config/ima/). The skill includes explicit instructions for the AI to never reveal, log, or echo these credentials during operation.
[EXTERNAL_DOWNLOADS]: Communicates with official IMA API endpoints (ima.qq.com) and Tencent Cloud Object Storage (myqcloud.com) for managing note data and uploading knowledge base documents. All external communications are restricted to well-known service infrastructure belonging to the vendor.
[COMMAND_EXECUTION]: Utilizes system utilities including curl, jq, python3, and node for API interaction, JSON parsing, and character encoding conversion (UTF-8). These operations are implemented as utility helpers for data processing and do not involve dynamic execution of untrusted code.
[DATA_EXFILTRATION]: Implements file upload functionality to synchronize user-selected documents to the IMA knowledge base. This data transfer is the intended primary purpose of the skill and is directed exclusively to official vendor-controlled storage.

ima-skill