The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill is designed to access and process a configuration file named siyuan.json which is documented to store sensitive authentication information, including a plaintext username and password for WebDAV access.
[COMMAND_EXECUTION]: The execution flow involves the agent calling subprocess.run to execute the scripts/upload_asset.py script with arguments, creating a surface for command injection if the input parameters are not strictly controlled.
[EXTERNAL_DOWNLOADS]: The skill requires the manual installation of the requests Python library via pip, which is used by the internal script to perform network operations.
[DATA_EXFILTRATION]: The skill transmits data externally to a WebDAV server specified in the configuration; while this is a core feature, it involves sending user-generated content and credentials over the network.
[PROMPT_INJECTION]: The skill includes indirect injection risk factors:
Ingestion points: The agent reads the siyuan.json file and processes user instructions to generate SVG content.
Boundary markers: There are no boundary markers or instructions to the agent to ignore potentially malicious content within the SVG data or the configuration file.
Capability inventory: The skill possesses the ability to write to the local file system and perform arbitrary HTTP PUT requests via the requests library in upload_asset.py.
Sanitization: While upload_asset.py performs basic character replacement on the file name (topic), it does not validate or sanitize the svg_content before writing it to disk or sending it to the remote server.

siyuan-excalidraw