AgentSkillsCN

human-skills-router

提升协作的包容性：营造心理安全感，促进更清晰的参与方式，推动更具意识的沟通，确保决策更加公平公正。当您在多元团队中开展工作、主持讨论、推进新人入职，或当团队中某些声音长期被忽视、紧张氛围看似无形却潜藏隐患时，此技能将助您打破壁垒，让每一位成员的声音都被听见、被重视。

SKILL.md
--- frontmatter
name: human-skills-router
description: Routes a human-skills request to the right coaching skill (feedback, conflict, stakeholder alignment, meeting facilitation, decision making, resilience, self-development, inclusion, ethical technology). Use when the user asks for “human skills”, communication help, difficult conversations, collaboration issues, growth coaching, stress, or responsible/ethical tech use.

Human Skills Router

Purpose

Ask minimal clarifying questions, select the right skill, and produce a useful first draft fast.

Available skills

Intake (ask max 3 questions)

  1. •Who is involved and what is the relationship? (peer, student, manager, client, team)
  2. •What outcome do you want after this? (decision, alignment, behavior change, repair, boundary, plan, reduce stress)
  3. •What constraint matters most? (time pressure, tension level, written vs live, formal vs informal, privacy/compliance)

If the user already provided enough context, skip questions and proceed.

Routing rules

  • •If the user mentions feedback, performance, review, “how do I say this”, “address behavior” → use feedback-coaching.
  • •If there is tension, blame, disagreement, frustration, escalation risk, “difficult conversation” → use conflict-navigation.
  • •If it’s expectations, scope creep, requirements, stakeholder updates, negotiating constraints → use stakeholder-alignment.
  • •If it’s a group setting: meeting/workshop/retro, stuck team, need agenda/decision process → use meeting-facilitation.
  • •If it’s choosing between options, prioritization, trade-offs, uncertainty, ownership → use decision-making.
  • •If the user is overwhelmed, stuck, stressed, anxious, needs boundaries or recovery → use resilience-and-self-regulation.
  • •If the user asks about growth, learning plan, portfolio evidence, strengths/weaknesses, performance prep → use career-self-development.
  • •If voices are missing, meeting dominance, inclusion, psychological safety, bias-aware collaboration → use inclusive-collaboration.
  • •If the question involves privacy, sensitive data, consent, transparency, bias, IP, analytics, AI/tooling risk → use ethical-technology-use.

Output contract (always deliver something)

Return:

  1. •Recommended skill (by name) + one-line reason.
  2. •First usable artifact (message draft / conversation script / agenda / decision record / 30-minute plan), using the target skill’s output defaults.

Defaults

  • •Provide two tones when drafting messages: (A) direct, (B) warmer.
  • •Keep it concrete: observable facts, explicit request, clear next step.
  • •Use the smallest effective intervention: fewer steps, less jargon.

Guardrails

Do not invent policies, agreements, or facts. If HR/legal/compliance risk is high, keep language factual and recommend consulting the right channel.