AgentSkillsCN

new-vendor-evaluation

生成供应商评估评分卡，以加权标准进行综合考量。当用户希望评估供应商、对比多家供应商，或对某款工具进行评估时，此技能将为您提供支持——评估维度包括安全性、合规性、成本，以及功能能力。

SKILL.md
--- frontmatter
name: "new-vendor-evaluation"
description: "Generates vendor evaluation scorecards with weighted criteria. Use when the user asks to evaluate a vendor, compare vendors, or assess a tool. Categories: security, compliance, cost, capability."
user-invocable: true
allowed-tools: ["Read", "Write", "Edit", "Glob", "Grep", "Bash", "Task"]

New Vendor Evaluation

Generate a vendor evaluation scorecard for Pivot Point Consulting.

Steps

  1. •

    Read proven patterns from .claude/feedback/patterns.md for established conventions and lessons learned.

  2. •

    Read the vendor evaluation template from templates/vendor-evaluation-template.md to understand the required document structure.

  3. •

    Read the style guide from marketing/brand/style-guide.md for voice, tone, and formatting standards.

  4. •

    Ask the user for the following inputs (if not already provided):

    • •Vendor name: The vendor or tool being evaluated.
    • •Evaluation context: Why this evaluation is being conducted (e.g., new tool selection, renewal decision, compliance audit).
    • •Evaluation categories: Which categories to prioritize (security, compliance, cost, capability, support, integration).
  5. •

    Check for duplicates by scanning vendors/evaluations/[year]/ for existing evaluations of the same vendor. Alert the user if one already exists.

  6. •

    Generate the vendor evaluation document with the following sections:

    • •Frontmatter (YAML):
      • •title: Evaluation title
      • •vendor: Vendor name
      • •product: Product or service being evaluated
      • •evaluator: [PLACEHOLDER]
      • •date: Current date
      • •status: draft
      • •decision: pending | approved | rejected | deferred
    • •Vendor Overview: Brief description of the vendor, their product/service, and market position.
    • •Evaluation Purpose: Why this evaluation is being conducted and what decision it will inform.
    • •Weighted Criteria Table: Markdown table with columns:
      • •Category
      • •Criterion
      • •Weight (%)
      • •Score (1-5): [PLACEHOLDER]
      • •Weighted Score: [PLACEHOLDER]
      • •Notes
    • •Scoring Rubric:
      • •5 = Exceeds requirements
      • •4 = Meets all requirements
      • •3 = Meets most requirements
      • •2 = Partially meets requirements
      • •1 = Does not meet requirements
    • •Category Assessments: For each evaluation category, include:
      • •Security: Data protection, encryption, access controls, incident response, SOC 2 compliance.
      • •Compliance: HIPAA, HITRUST, regulatory adherence, audit history.
      • •Cost: Licensing model, total cost of ownership, hidden costs, ROI projection.
      • •Capability: Feature completeness, scalability, performance, roadmap alignment.
      • •Support: SLAs, responsiveness, documentation quality, training resources.
      • •Integration: API availability, interoperability, migration effort.
    • •Assessment Summary: Overall score, strengths, weaknesses, and risks.
    • •Recommendation: Clear recommendation with rationale (Approve / Reject / Defer / Further Evaluation Needed).
    • •Alternatives Considered: Brief notes on other vendors or approaches evaluated.
  7. •

    Write the file to vendors/evaluations/[year]/[kebab-case-vendor]-evaluation.md where [year] is the current year.

  8. •

    Update the directory README at vendors/evaluations/[year]/README.md by adding the new evaluation to the index list. If the README does not exist, create it with a header and the first entry.

  9. •

    Log the action by appending a JSON line to .claude/feedback/skills-log.jsonl with:

    json
    {"skill": "new-vendor-evaluation", "action": "created", "path": "<file-path>", "timestamp": "<ISO-8601>", "vendor": "<vendor-name>", "decision": "pending"}