本文へ移動
cccskills
無料GitHub で公開

cowork-vendor-comparison

Drop a folder of vendor quotes, proposals, and SOWs on Cowork -- normalizes them into one comparison matrix, computes true total cost of ownership, surfaces the terms each vendor buried, and arms you with negotiation leverage points and reference-check questions.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md3.1 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Cowork Vendor Comparison

Compare vendor proposals the way a procurement lead does: normalize everything to the same units before judging anything, price the full term not the first invoice, and read the terms the sales deck skipped. Input: a folder of quotes/proposals/SOWs (PDF, .docx, .xlsx) and, ideally, one sentence on what the purchase needs to accomplish.

Workflow

  1. Extract per vendor: pricing structure (license, usage, implementation, support tiers), contract term and renewal terms, SLAs and remedies, implementation timeline and dependencies, what is in scope vs. priced as add-on, exit terms (data export, termination fees), and every assumption the proposal states.
  2. Normalize. Convert all pricing to a common basis -- same seat count, same usage volume, same term. Where a proposal is silent on something another proposal prices (training, integrations, overage rates), mark it UNPRICED -- ask, never zero. Unpriced is not free.
  3. Compute TCO over the realistic term (default 3 years): year-one cost, steady-state annual cost, escalators applied, one-time costs amortized, and the exit cost. Show the math per vendor.
  4. Matrix. Output vendor-comparison.md: side-by-side table of cost, capability fit against the stated need, SLA strength, implementation risk, and contract flexibility -- each cell citing the source document and page. Follow with a plain-English paragraph per vendor: the honest case for and against.
  5. Negotiation prep. For the top 2 vendors: leverage points (their weaknesses vs. the rival's strengths, end-of-quarter timing, multi-year vs. flexibility trades), the specific asks worth making (cap the escalator, free implementation, opt-out at 12 months), and 5 reference-check questions targeting each vendor's specific risk areas.

Rules

  • Never judge from unnormalized numbers. A lower sticker with usage pricing can be the expensive option.
  • Every cell in the matrix traces to a document and page. "Vendor B has better support" without a citation is vibes.
  • Silence is a finding: what a proposal does not say (no uptime SLA, no data-export clause) goes in the matrix as prominently as what it says.
  • Do not manufacture a winner. If the vendors genuinely split on cost vs. capability, present the trade and the decision criteria, and say which way the stated need leans.
  • Flag lock-in explicitly: proprietary data formats, migration fees, auto-renewals with long notice windows.
  • If proposals answer different scopes (one quoted 50 seats, one quoted 200), lead with that mismatch and request aligned quotes before deep comparison.

Quick Commands

  • "Compare [folder]" -- full workflow
  • "TCO only" -- steps 1-3
  • "What's unpriced?" -- the ask-list of gaps per vendor
  • "Prep me to negotiate with [vendor]" -- step 5 for one vendor

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Audit websites for accessibility issues and WCAG compliance. Use when checking accessibility, fixing a11y issues, or ensuring WCAG compliance.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Deploy a 2-layer parallel agent hierarchy for large, parallelizable work — big refactors, multi-file migrations, codebase-wide audits, bulk generation. A top-tier commander (Fable or Opus) orchestrates the swarms; the user picks a power level (Max Power / Heavy / Balanced / Economy) that sets the Opus/Sonnet/Haiku model mix per layer. Layer 1 is 3-50+ specialist agents, each with its own full context window; Layer 2 is 2+ sub-agents per member. Includes git safety, tiered sizing, a pre-deploy gate, phantom-completion checks, and multi-wave follow-up.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Use to test whether an AI agent, prompt, or skill actually works before or after it ships — build a small eval set, score runs, and catch regressions when the prompt or model changes. Trigger when the user asks whether a prompt is good, wants to compare models or versions, sees inconsistent agent output, or is about to put an agent in front of customers.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Deploys swarms of sub-agents for massive parallel data processing tasks. Unlike agent-army (which is for code changes), this is for DATA tasks -- processing 1000 documents, analyzing datasets, bulk content generation. Configurable swarm size, task distribution, result aggregation, progress tracking, and error recovery.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Designs and deploys custom agent teams for specific business workflows. Interactive discovery of business processes, then generates complete team configurations with specialized agent roles, tool access, communication protocols, and handoff rules.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

Agent-to-Agent (A2A) communication protocol. Connect two or more Claude agents that pass messages, share context, delegate tasks, and collaborate. Implements structured handoffs, shared memory, and multi-agent conversations.

日本語の概要は準備中です。原文の説明を表示しています。

OneWave-AI/claude-skills3392026年10月2日 更新

OneWave-AI のスキルをすべて見る

このスキルの問題を報告する