The Vendor Desk
Skills · Subagent · Command · PDF
.claude/the-vendor-desk/zip
skills/06
├ exit-checklist
├ pilot-design
├ pricing-trap-review
├ vendor-comparison
├ vendor-evaluation
├ workflow-baseline
agents/01
├ vendor-doc-reader
commands/01
├ /evaluate
FIELD-GUIDE.pdf9pp
English
One payment · No account · No subscription
- ✓ Download link in your inbox in seconds
- ✓ Secure card checkout · VAT handled
- ✓ 14-day money back, no questions
- ✓ Every future revision included
- ✓ Runs in Claude Code and claude.ai
Built for Claude first, our recommended agent. Skills follow the open Agent Skills standard (SKILL.md), so they also run in any agent that reads it; subagents and slash commands are Claude Code formats.
The Vendor Desk
The AI buying decision, run as a test
six skills·one subagent·one command·9-page guide
The AI tool on the table, scored against the workflow it would enter: seven questions answered from evidence, a pilot with a threshold, and the cost at ten times the volume.
Executive summary
- The problem
- An AI tool is usually bought on the vendor's demo and benchmarks, against a workflow nobody wrote down, so there is nothing to hold the pilot against. The pilot never fails, it just never ends, and the bill arrives at production volume.
- What it does
- Six skills and a read-only vendor-doc-reader write the workflow baseline first, score each vendor on seven questions from evidence, design a four-week pilot with a threshold and kill criteria, model the cost at real volume and check the exit.
- Why take it
- Your name goes on the budget line. This makes the call to buy, build or wait one you can show your working for, and reverse if the tool does not deliver.
For the person who has to decide whether an AI tool earns a line in next year's budget.
In the first week you will
- 01Write the five-field workflow baseline with workflow-baseline before a single vendor is scored
- 02Run /evaluate on the vendor from last week and get the seven-question scorecard with a build, buy or wait call
- 03Hand the contract to the vendor-doc-reader subagent and get what it says on each question and where it is silent
Contents of the Kit
- 01Six installable agent skills: workflow-baseline, vendor-evaluation, pilot-design, pricing-trap-review, exit-checklist, vendor-comparison
- 02The vendor-doc-reader subagent: reads contracts, security questionnaires and product documentation, and reports what they say and what they do not say. Read-only
- 03An /evaluate slash command that opens a scorecard for a named vendor from one sentence
- 04The seven-question buying test as a scorecard with the evidence per question supplied by you, and a pilot design with a baseline, a threshold and kill criteria written before week one
- 05Blank templates for the workflow baseline and the vendor scorecard
- 06A 9-page Field Guide on buying against a documented workflow, and plugin manifest included
One skill, as it ships
SKILL.md
workflow-baseline
- Use when
- Use when someone wants to evaluate, pilot or buy an AI tool, when a demo has happened and a budget exists, when a vendor offers a trial, or when someone asks "should we buy this". Refuses to open an evaluation without the five fields and the baseline.
- Requires
- The workflow, in the user's words · The five fields · The thin baseline
- Will not
- No evaluation opens without the five fields and a run baseline · Never estimate hours, volumes, rates or error rates · Do not evaluate the vendor here
- Returns
- Workflow baseline: {workflow name}
One of six skills in the kit · All of them install the same way
The contents, in full · six skills, one subagent, one command
- 01workflow-baselineThe five fields the buying decision rests on: who does the work today, how often, how long it takes, what it costs, and what happens when it goes wrong, plus the thin baseline (what the current best alternative produces, at what quality, for what cost). Refuses to open an evaluation without one.
- 02vendor-evaluationThe seven-question scorecard for one vendor: failure mode and how you would know, what happens to your data, what it replaces, the cost at ten times the usage, the setup before it is useful, whether you can leave, and the thin baseline. Each question is scored from the evidence you supply, with the six hype markers checked and the build, buy or wait call at the end.
- 03pilot-designFour weeks, not six months: scope, the one primary metric and the pass threshold written in week zero, the evaluation set of twenty to fifty real examples, tool and baseline run in parallel, the kill criteria, the owner, and the decision record with its reversal trigger.
- 04pricing-trap-reviewSeat, usage and outcome pricing side by side for one price sheet: minimums, overage, what the pilot volume hides, what the renewal will cost, what happens to the price when the underlying model changes. Every projection is arithmetic on your figures, labeled as an assumption.
- 05exit-checklistBefore signing: what can be exported and in what format, what is deleted and when, the contract term and notice period, where the configuration lives, and what breaks in your operation the day the tool is switched off.
- 06vendor-comparisonTwo or more completed scorecards side by side, weighted by criteria you set, with the disagreements between them shown rather than averaged away, and the questions still open for each vendor.
- 07vendor-doc-reader (subagent)Reads a contract, a security questionnaire, a data processing agreement or the product documentation, and reports what the document says on each of the seven questions and where it is silent, with the clause quoted. It reads; it never fills a gap with an assumption.
- 08/evaluateThe slash command that opens the seven-question scorecard for a named vendor from one sentence.
The Age of Coexistence with AI names the trap: pilots are cheap and scaling is expensive, so the rational local decision is always to fund another pilot, and an organization that cannot evaluate a system on its own buys on the vendor’s testimony. The Vendor Desk is the job after the demo: the budget exists, and somebody has to decide whether this tool earns a line in next year’s costs. Accuracy claims without a baseline are recorded as claims, and every projection is arithmetic on figures you supply.
Pairs with
The Age of Coexistence with AI explains why the organizations that capture value buy workflow redesign and not licences, and why the proof-of-concept trap keeps the rest funding pilots. The journal essay on evaluating AI tools without the hype is the buying process this kit runs. The Coexistence Desk decides which workflows to reinvent; the Vendor Desk decides which tools, if any, to buy for them.
Before you buy
01Do I need Claude Code?
No. Every skill uploads to claude.ai as a zip. Subagents and slash commands are the Claude Code extras.
02I am not a programmer.
Install is copying folders into a .claude directory. The Field Guide in each kit assumes nothing beyond that.
03Will it work in my agent?
Skills follow the open Agent Skills standard (SKILL.md), so any agent that reads it can run them. Built for Claude first, our recommended agent.
04Do the tools invent numbers?
Never. Every figure is supplied by you, confirmed by you, or labeled as an assumption in the output.
05How does it arrive?
Instantly. Pay by card and the download link is in your inbox within seconds, and on the confirmation screen.
06Is this a subscription?
No. One payment, lifetime access to your download, every future revision of what you bought included.
07What if it is not useful to me?
Write to info@conclusionmp.com within 14 days of purchase and you get a full refund. No questions, no forms.
Ready when you are
The Vendor Desk, $12.
Instant download, one payment, every future revision included. Or take The AI Shelf, six items for $39, and this kit works out at $6.50.