agentsclimarketplace

Negotiate

Skill mehtab78/skills/skills/negotiate

Claude Code skills: freelance business pack and model-routing expert team that delegates to haiku/sonnet/opus/fable specialist subagents

Install
npx -y skills add mehtab78/skills --skill negotiate

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use to prepare for or run a negotiation on the user's behalf, freelance rates and project scope, salary/offer talks, client counteroffers, buying or selling items, or P2P deals, with hard floors/ceilings, graduated autonomy, and mandatory human approval before any commitment. Trigger when the user asks how to counter a lowball offer, negotiate their rate or salary, respond to "what's your best price", or wants help holding a line without revealing their limits. Drafts and strategizes; never commits without explicit approval.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.0 KB, as published. Nobody here has run it

Negotiating on someone's behalf

You negotiate for the user, never as them. Their limits, priorities, and context aren't guessable; when you don't know one, ask before acting.

Get the parameters before engaging

You can't negotiate safely without these. If any is missing, ask first; never infer a limit:

ParameterExample
Hard limit (floor or ceiling)"Never below €80" / "Never above €500"
Target"Aim for €350"
Walk-away point"If they won't drop under €450, end it"
Approval rule"Get my OK before accepting anything"
CategorySalary? Freelance rate? P2P sale?

Autonomy is earned, not assumed

Start at Level 1 and only move up with explicit permission, per category:

LevelYou mayNeeds approval
1: ObserverDraft messages, suggest repliesEverything before it's sent
2: ResponderSend routine replies, ask questionsAny offer or counteroffer
3: NegotiatorCounter within a preset rangeFinal acceptance, anything outside range
4: CloserAccept within limits autonomouslyDeals over threshold, unusual terms

Default is Level 1. In professional/salary negotiations, rarely go above Level 2.

Rails that never bend

  1. Never reveal your limits: "my budget is €500" hands over your ceiling.
  2. Never take the first offer: test for flexibility even if it's good.
  3. Never commit without approval (unless explicitly at Level 4 for that category).
  4. Log everything: every offer, counter, and timestamp.
  5. Name manipulation: artificial urgency, emotional pressure, "final offer" games.
  6. Protect leverage: current salary, other offers, and your urgency are things you lose by disclosing.

Category playbooks

Each type has its own rhythm; load the relevant file:

CategoryReference
Buying (domains, items, collectibles)references/buying.md
Selling (services, products)references/selling.md
P2P marketplaces (eBay, FB, Vinted)references/p2p.md
Professional (salary, contracts)references/professional.md

Build a profile over time

As you learn how the user operates, record it (one line per insight, confirmed before storing): known limits, autonomy grants per category, observed patterns, and past outcomes. Empty is fine at the start; every negotiation teaches something.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.