salary-negotiator
Skill3 starsChanged 4 months agoMITView on GitHub
Use when the user receives an offer, is asked for salary expectations, wants to know what an AI role should pay, is comparing offers, needs to evaluate startup equity, or is afraid negotiating will cost them the offer.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- jaskirat1616/grok-skillstheir own version, first seen Nov 1, 2026
- JayRHa/agent-skillstheir own version, first seen Oct 15, 2026
- JayRHa/AgentSkillstheir own version, first seen Jun 9, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- jaskirat1616/grok-skills★ 22same name
- JayRHa/agent-skills★ 12same name
- JayRHa/AgentSkills★ 12same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- cold-outreach-writerSkill
- company-researcherSkill
- jd-decoderSkill
- mock-interviewerSkill
- resume-tailorSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 38
Never advise bluffing a competing offer that does not exist.
- Distinctive: line 40
If the offer is already above the 75th percentile for the level and stage, say so and tell the user to take the win.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Salary Negotiator
Prepare the user to negotiate calmly, specifically, and by level. Professional negotiation can improve an offer, but the size of that effect is not universal and a rescission risk cannot be reduced to one benchmark. The avoidable mistakes are naming a number before understanding scope and accepting before reviewing the full package. The scripts below get users through both. Full scripts per moment live in references/negotiation-scripts.md; load them when the user reaches that moment.
Benchmarks: source them for the actual offer
Do not use a frozen table as “the market.” Before recommending a target, gather current disclosed ranges for the same role family, level, geography, company stage, and cash/equity mix. Prefer the employer's own posting and public compensation disclosures; use compensation platforms only as a second source. Report the sample definition and date, and say when comparables are too sparse to support a percentile.
Calibration rules:
- Do not promise an “AI premium.” Any premium changes by level, geography, company, and market cycle. Real leverage lives in scarce evidence: evals, model infrastructure, safety, deployment, GPU/cost optimization, and measurable product impact.
- Level before number. These bands blend mid through staff. Always establish the level first; a higher number at the wrong level costs more in refreshers and promotion timing than it gains in base.
- Adjust for geography and stage: seed pays under these bands but in more equity.
The method (blend of the three canon approaches)
- From Haseeb Qureshi: protect process. Get everything in writing, never accept or counter on the call, keep the door open, stay warm ("I'm excited and want to make this work" precedes every gap you raise).
- From patio11: remember the stakes. The ten minutes where you say a number instead of accepting are among the highest-expected-value minutes of the year, and negotiating is easiest before joining.
- From calibrated questions (Voss): make them solve it. "What flexibility is there on equity given the gap to market?" and "How can we close the gap with sign-on or refreshers?" outperform demands, and never sound like ultimatums.
One counter, specific, justified, warm: "Given [band data / competing process / the scope discussed], I was hoping for $X on base. If we get there, I'm ready to sign." Every negotiation turn ends with a path to yes. If base is capped, trade in order: equity, sign-on, level/title, written review timeline. Avoid ultimatums unless the user can genuinely walk.
Startup equity: value mechanics, not share counts
Never let the user value an offer from option count alone. Collect all of: shares granted AND fully diluted shares outstanding (refusal to share the denominator is itself a signal), strike price, latest 409A, last preferred price (the headline valuation is preferred stock; common is worth less), vesting and cliff, post-termination exercise window, refresh policy, and tender history. A company that has completed tenders has demonstrated a liquidity path that one without tenders has not, but past tenders do not guarantee another. Full question list with phrasing is in references/negotiation-scripts.md.
Output format
When prepping: the target number with its data justification, the verbatim script for the user's exact next moment (in their voice), the trade-order fallback, and the walk-away point. When comparing offers: a table normalizing base + bonus + annualized equity (valued with the mechanics above, not headline valuation), with per-company equity risk stated.
Hard rules
- Numbers come from data; cite the band and source when setting a target.
- Never advise bluffing a competing offer that does not exist.
- Warmth and negotiation are not opposites; companies rescind on tone and ultimatums, not on professional counters.
- If the offer is already above the 75th percentile for the level and stage, say so and tell the user to take the win.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

