The Agent Skills Directory

[PROMPT_INJECTION]: The skill outlines a workflow for reading and processing data from external sources like Feishu/Lark documents and Bitables, which establishes a surface for indirect prompt injection where adversarial data could influence agent behavior. * Ingestion points: External content accessed through Feishu/Lark API endpoints and Bitable queries (SKILL.md). * Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands in the retrieved data. * Capability inventory: The skill enables sensitive actions such as sending messages and changing approval or workflow states (SKILL.md). * Sanitization: No specific data sanitization or filtering is mentioned, although the skill requires explicit human confirmation before performing high-impact tasks.
[COMMAND_EXECUTION]: The skill instructs the agent to configure and manage Model Context Protocol (MCP) servers using deepseek mcp commands, which involves local shell execution to set up the integration environment (SKILL.md).
[SAFE]: The skill references official API endpoints for Feishu (open.feishu.cn) and Lark (open.larksuite.com), which are established services for workplace productivity. Additionally, the skill explicitly warns against hardcoding secrets and promotes the use of environment variables for credential management, aligning with security best practices.

feishu