← 返回 Skills 市场
692
总下载
2
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install negotiate
功能描述
Negotiate on behalf of your principal with hard limits, graduated autonomy, and mandatory human approval for commitments.
安全使用建议
This skill appears to do what it claims (a negotiation assistant) and doesn't request extra credentials or install anything. Before installing, decide and document: (1) where logs and the negotiation profile will be stored, who can access them, and how long they're retained; (2) the exact mechanism by which the agent will send messages or make commitments (ensure human approval gates are enforced); and (3) policy for handling sensitive inputs (current salary, bank/payment details) so the agent doesn't accidentally record or transmit them. If you will let the agent act autonomously at higher levels, explicitly record per-category autonomy grants and review them periodically.
功能分析
Type: OpenClaw Skill
Name: negotiate
Version: 1.0.0
The OpenClaw AgentSkills bundle 'negotiate' is designed with strong safety mechanisms and explicit instructions to limit agent autonomy and prevent unauthorized actions. The `SKILL.md` file, which serves as the agent's instructions, repeatedly emphasizes mandatory human approval, default low autonomy levels, explicit parameter requirements, and logging of all actions. There are no indicators of data exfiltration, malicious execution, persistence, obfuscation, or prompt injection attempts designed to subvert the agent's intended purpose or bypass security controls. The included negotiation strategy files (`buying.md`, `p2p.md`, `professional.md`, `selling.md`) consistently reinforce escalation triggers requiring principal approval for sensitive actions.
能力评估
Purpose & Capability
Name, description, and included docs (buying.md, selling.md, p2p.md, professional.md) all align with a negotiation assistant. The skill requests no environment variables, binaries, or other resources that would be unrelated to negotiation functionality.
Instruction Scope
Runtime instructions are specific about required negotiation parameters, autonomy levels, and escalation triggers. They instruct the agent to 'Log everything' and to 'Build' a negotiation profile over time but do not specify where or how to store these logs/profiles or what retention/permission model to use—this is a deliberate design choice but leaves an operational decision to the agent environment.
Install Mechanism
Instruction-only skill with no install spec and no code files to execute. Low risk from install mechanism because nothing is downloaded or written by the skill itself.
Credentials
The skill declares no required environment variables, credentials, or config paths. It does not ask for unrelated secrets or system access in its instructions.
Persistence & Privilege
always:false and autonomous invocation allowed (default) are appropriate. The only persistence-related concern is that the instructions encourage building and storing a negotiation profile and 'logging everything'—those are reasonable for this purpose but the skill does not define storage boundaries, access controls, or data retention, which are left to the host environment.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install negotiate - 安装完成后，直接呼叫该 Skill 的名称或使用
/negotiate触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
Initial release
元数据
常见问题
Negotiation 是什么？
Negotiate on behalf of your principal with hard limits, graduated autonomy, and mandatory human approval for commitments. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 692 次。
如何安装 Negotiation？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install negotiate」即可一键安装，无需额外配置。
Negotiation 是免费的吗？
是的，Negotiation 完全免费（开源免费），可自由下载、安装和使用。
Negotiation 支持哪些平台？
Negotiation 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Negotiation？
由 Iván（@ivangdavila）开发并维护，当前版本 v1.0.0。
推荐 Skills