The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, credential exfiltration, or unauthorized persistence mechanisms were detected in the skill instructions.
[COMMAND_EXECUTION]: The skill utilizes lark-cli to perform operations against the Lark API. This is a vendor-provided tool consistent with the skill's purpose for interacting with Lark services.
[DATA_EXPOSURE]: No hardcoded secrets, API keys, or access to sensitive filesystem locations (e.g., ~/.ssh, ~/.aws) were found. Authentication is handled via a shared configuration referenced in the instructions.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes external approval task data and possesses write capabilities (approve/reject).
Ingestion points: instances.get, tasks.query (SKILL.md)
Boundary markers: None specified in this file
Capability inventory: lark-cli approval tasks [approve|reject|transfer], lark-cli approval instances [cancel|cc] (SKILL.md)
Sanitization: No explicit sanitization or validation of the retrieved task content is mentioned.

lark-approval