The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill's 'register' functionality encourages the storage of sensitive 'session keys' and 'App IDs' in a Feishu document. Storing authentication tokens in collaborative docx documents can lead to unauthorized access or credential theft if the document permissions are not strictly managed.
[DATA_EXFILTRATION]: Sensitive robot credentials (session keys) are transmitted to and stored on Feishu's cloud infrastructure as part of the registry functionality.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection.
Ingestion points: The listRobots function in index.js retrieves arbitrary content from a Feishu document.
Boundary markers: There are no markers or instructions to isolate the retrieved document content from the agent's context.
Capability inventory: The skill allows creating and updating documents, and it interfaces with the broader agent environment.
Sanitization: No validation or escaping is performed on the data fetched from the registry document, allowing potentially malicious instructions to be processed by the agent.
[COMMAND_EXECUTION]: The skill uses the commander library to handle command-line arguments. While no direct shell execution of these arguments was observed, user-supplied strings (like robot names) are interpolated into the content sent to the Feishu API.
[EXTERNAL_DOWNLOADS]: The skill communicates with Feishu's official Open APIs (open.feishu.cn) to perform search, create, and update operations on documents.

feishu-robot-registry