AgentSkillsCN

stakeholder-discovery-template

用于创建新项目特定利益相关者技能的通用利益相关者发现模板。当用户请求创建利益相关者技能、利益相关者分析框架或团队上下文模板时自动触发。不要在实际利益相关者讨论期间加载（改用项目特定技能）。

SKILL.md
--- frontmatter
name: stakeholder-discovery-template
description: Generic stakeholder discovery template for creating new project-specific stakeholder skills. Auto-invoke when user requests stakeholder skill creation, stakeholder analysis framework, or team context templates. Do NOT load during actual stakeholder discussions (use project-specific skills instead).
allowed-tools: [Grep]
version: 1.0.0
category: Templates
tags: [template, stakeholder-discovery, skill-creation]
last-updated: 2025-10-19

Stakeholder Discovery Template

Purpose

This template guides the creation of project-specific stakeholder skills. Replace all [PLACEHOLDER] values with actual project information.

SKILL.md Frontmatter Template

yaml
---
name: [project]-[domain]-stakeholders
description: Stakeholder context for [PROJECT_NAME] [DOMAIN] when discussing [USE_CASE_1], [USE_CASE_2], or [USE_CASE_3]. Auto-invoke when user mentions [PROJECT_TRIGGER], [DOMAIN_TRIGGER], or [TEAM_TRIGGER]. Do NOT load for general [DOMAIN] discussions unrelated to [PROJECT_NAME].
allowed-tools: []
version: 1.0.0
category: Stakeholders
tags: [[project], [domain], [key-theme-1], [key-theme-2]]
last-updated: [YYYY-MM-DD]
---

Description Engineering Guidance:

✅ DO include specific trigger terms:

  • •Project name (exact capitalization users will mention)
  • •Domain keywords (fintech, UX research, payment processing)
  • •Team names users will reference

❌ DON'T use generic descriptions:

  • •"Provides stakeholder information" (too broad)
  • •"Use when discussing stakeholders" (will load too often)
  • •No explicit project scope (will contaminate other projects)

Example Good Description:

code
Stakeholder context for Phoenix UX research project when discussing user testing, research synthesis, or design validation. Auto-invoke when user mentions Phoenix, UX research stakeholders, or design team collaboration. Do NOT load for general UX discussions unrelated to Phoenix.

Skill Content Structure Template

[TEAM_NAME] Team

Team Objectives

  • •[PRIMARY_OBJECTIVE_1] (quantified if possible: >95% accuracy, <100ms latency)
  • •[PRIMARY_OBJECTIVE_2]
  • •[PRIMARY_OBJECTIVE_3]

Key Stakeholders

[ROLE_TITLE] (Use role, not individual name - NO PII)

  • •Decision Authority: [SPECIFIC_DECISIONS_THIS_ROLE_MAKES]
  • •Communication Preference: [PREFERRED_CHANNELS_AND_STYLE]
    • •Examples: "Data-driven with A/B test results", "Formal written proposals", "Technical RFCs with diagrams"
  • •Success Metrics: [QUANTIFIABLE_METRICS_THIS_ROLE_CARES_ABOUT]
    • •Format: "[Metric name] [operator][value]" (e.g., "Fraud detection rate >95%")
  • •Pain Points:
    • •[CURRENT_CHALLENGE_1] (quantified if possible)
    • •[CURRENT_CHALLENGE_2]
    • •[CURRENT_CHALLENGE_3]

Decision-Making Patterns

  • •[DECISION_TYPE]: [APPROVAL_PROCESS_AND_CRITERIA]
    • •Example: "Rule Changes: Requires statistical significance (p<0.05) with 2-week A/B test"
  • •Review Cycle: [FREQUENCY_AND_FORMAT]
    • •Example: "Weekly sprint reviews for non-critical updates"
  • •Escalation Protocols: [WHEN_AND_HOW_TO_ESCALATE]
    • •Example: "Emergency fraud spike triggers immediate team call"
  • •Approval Chain: [ROLE_1] → [ROLE_2] → [ROLE_3] (for [DECISION_SCOPE])

Communication Channels (Optional)

  • •Primary: [SLACK_CHANNEL_OR_EMAIL_GROUP]
  • •Meetings: [RECURRING_MEETING_SCHEDULE]
  • •Documentation: [CONFLUENCE_WIKI_OR_DOC_LOCATION]
  • •Escalations: [EMERGENCY_CONTACT_METHOD] (NO personal phone numbers)

Validation Checklist

Before finalizing skill, verify:

Content Quality:

  • • All [PLACEHOLDERS] replaced with actual values
  • • Zero PII (no names, emails, phone numbers, personal identifiers)
  • • Zero business-confidential metrics (use relative: ">95%" not "$5M revenue")
  • • Quantified metrics where possible (not "high accuracy", say ">95%")

Description Engineering:

  • • Description includes specific WHEN triggers (project name, domain keywords)
  • • Description includes explicit WHEN NOT boundaries
  • • Tested with positive queries (should load): [LIST_TEST_QUERIES]
  • • Tested with negative queries (should NOT load): [LIST_TEST_QUERIES]

Metadata & Tracking:

  • • Ground truth source documented with date
  • • Validation date included
  • • Next review date specified (recommend quarterly)
  • • allowed-tools: [] set (stakeholder skills are read-only reference)
  • • Category and tags aid discovery

Security & Scope:

  • • File location matches scope (project skills in .claude/skills/projects/[name]/)
  • • No personal preferences mixed with stakeholder context
  • • Reviewed by at least one other team member (if team skill)

Usage Examples

Creating New Stakeholder Skill:

code
User: "Create a stakeholder skill for the Phoenix UX research team"

Claude: [Loads this template skill, uses structure to create Phoenix-specific skill]

Finding Template for Reference:

code
User: "What's the standard format for stakeholder skills?"

Claude: [Loads this template skill, shows structure and best practices]

Do NOT Use This Template For:

  • •Actual stakeholder discussions (use project-specific skills instead)
  • •Ground truth documentation (use ground-truth template)
  • •Product context (use product skill template)

Template Source: Research report + eval committee consensus Template Version: 1.0.0 Last Updated: 2025-10-19 Validation: Ready for Week 1 PoC testing