The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override safety guidelines or bypass instructions were found in the skill text.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a router that ingests untrusted user tasks and interpolates them into commands for sub-agents via message() and sessions_spawn(). This is an inherent part of its multi-agent coordination design.
Ingestion points: User-provided tasks described in the '当使用此技能' section.
Boundary markers: No delimiters or explicit instructions to ignore embedded commands are used when forwarding tasks.
Capability inventory: The skill utilizes inter-agent messaging and session spawning to execute logic across various domains (ops, finance, code).
Sanitization: No sanitization or validation of the input task string is performed before it is passed to sub-agents.

team-coordinator