The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install a global Node.js package @qianwenai/qianwen-cli using npm install -g. This package is a vendor-owned resource originating from the skill author, qianwen-ai. It is used to interact with the conference API.
[COMMAND_EXECUTION]: The skill performs shell operations to check for the presence of the qianwen CLI, verify its version (requiring >= 1.7.0), and manage authentication status via qianwen auth status. It also executes various subcommands of qianwen yunqi to fetch conference data.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from several sources, including the JSON output of the qianwen CLI and multiple local Markdown files (云栖大会会刊.md, 云栖大会参会指南.md, 数贸会语料.md). This creates a vulnerability surface where malicious instructions embedded in the conference agenda or guides could influence the agent's behavior.
Ingestion points: CLI output from qianwen yunqi list forums/exhibitors and local markdown reference files.
Boundary markers: Not present in the instruction set.
Capability inventory: Shell command execution and global npm package installation.
Sanitization: No specific sanitization or filtering logic is defined for the ingested data.

yunqi-agent-discovery