The Agent Skills Directory

[DATA_EXPOSURE]: The skill facilitates the management of the TEAMBITION_USER_TOKEN. It supports providing this token via environment variables or a local configuration file (user-token.json). While local storage in plain text is a common pattern for such tools, it represents a standard risk of credential exposure if the project directory is improperly secured or accidentally shared.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from the Teambition platform, including task titles, comments, and project descriptions.
Ingestion points: Data enters the agent context through scripts like query_tasks.py, query_task_detail.py, and render_rtf.py, which fetches and parses task notes and custom fields.
Boundary markers: The skill relies on natural separation in its output but does not implement explicit cryptographic boundary markers for the processed data.
Capability inventory: The skill's capabilities are limited to making authenticated HTTP requests to the Teambition API (open.teambition.com). It does not perform arbitrary shell command execution or dynamic code evaluation on ingested data.
Sanitization: The render_rtf.py script performs basic sanitization by using regular expressions to strip HTML tags and styles from rich text content before presenting it to the agent.
[EXTERNAL_DOWNLOADS]: The skill downloads project-related files and attachments from official Teambition/Alibaba Cloud OSS domains. These operations are essential to the skill's primary purpose and target trusted service endpoints.

dingtalk-teambition