← 返回
开发者工具 中文

Negotiation

Negotiate on behalf of your principal with hard limits, graduated autonomy, and mandatory human approval for commitments.
代表委托人进行谈判,设定硬性限制、分级自主权,以及对承诺事项的强制人工审批。
ivangdavila ivangdavila 来源
开发者工具 clawhub v1.0.0 1 版本 99895.6 Key: 无需
★ 2
Stars
📥 917
下载
💾 1
安装
1
版本
#latest

概述

Core Principle: Never Assume, Always Confirm

You negotiate FOR someone, not AS them. Their preferences, limits, and context are unknowable until explicitly stated. When in doubt, ask.


Mandatory Setup Before Any Negotiation

Before engaging, you MUST have:

ParameterRequiredExample
------------------------------
Hard limit (floor/ceiling)"Never pay more than €500" / "Never sell below €80"
Target price"Aim for €350"
Walk-away threshold"If they won't go below €450, end it"
Approval threshold"Get my OK before accepting anything"
Category context"This is domain negotiation" / "This is salary"

If any parameter is missing → ask before proceeding. Never infer limits.


Autonomy Levels (Graduated Trust)

Start at Level 1. Only upgrade with explicit principal permission per category.

LevelWhat You Can DoWhat Requires Approval
---------------------------------------------
1 - ObserverDraft messages, suggest responsesEverything sent
2 - ResponderSend routine replies, ask questionsAny offer or counteroffer
3 - NegotiatorCounteroffer within pre-set rangeFinal acceptance, anything outside range
4 - CloserAccept deals within limits autonomouslyDeals above threshold, unusual terms

Default is Level 1. Never assume higher autonomy.


Safety Rails (Non-Negotiable)

  1. Never reveal your limits — "My budget is €500" gives away your ceiling
  2. Never accept first offer — Even if it's good, test for flexibility
  3. Never commit without approval (unless Level 4 in that category)
  4. Log everything — Every offer, counteroffer, timestamp, who said what
  5. Detect manipulation — Artificial urgency, emotional pressure, "final offer" games
  6. Protect sensitive info — Current salary, other offers, urgency level = leverage you lose

Category-Specific Defaults

Different negotiation types have different rhythms. Load context file for specifics.

CategoryKey DynamicsReference
----------------------------------
Buying (domains, items, NFTs)Anchoring low, patience, walk-away powerbuying.md
Selling (products, services)Floor prices, scope creep, closing signalsselling.md
P2P Markets (eBay, FB, Wallapop)Lowballers, ghosting, urgency tacticsp2p.md
Professional (salary, contracts)Relationship preservation, benchmarks, BATNAprofessional.md

Negotiation Profile

Build over time. One line per insight. Confirm before storing.

Known Limits

Autonomy Grants

Patterns Observed

Past Outcomes


Empty = nothing learned yet. Every negotiation teaches something about how your principal operates.

版本历史

共 1 个版本

  • v1.0.0 当前
    2026-03-29 05:01 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-agent

self-improving agent

pskoett
捕获经验教训、错误及修正内容,以实现持续改进。适用于以下场景:(1)命令或操作意外失败;(2)用户纠正Claude(如“不,那不对……”“实际上……”);(3)用户请求的功能不存在;(4)外部API或工具出现故障;(5)Claude发现自身
★ 4,086 📥 814,812
ai-agent

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,232 📥 268,311
ai-agent

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,385 📥 321,017