← 返回 Skills 市场
Self-Improving Negotiation
作者
José I. O.
· GitHub ↗
· v1.0.0
· MIT-0
83
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install self-improving-negotiation
功能描述
Captures negotiation strategy failures, concession leaks, BATNA weakness, framing misses, objection handling gaps, escalation misalignment, anchor errors, an...
安全使用建议
This skill appears to do what it says: emit reminders and log negotiation learnings locally. Before enabling hooks or running scripts: (1) inspect the included scripts (activator.sh, error-detector.sh, extract-skill.sh) to confirm you are comfortable with their file-write behavior; (2) be aware the error-detector reads CLAUDE_TOOL_OUTPUT — avoid running it where tool output might contain secrets; (3) the hooks will add a virtual reminder at agent bootstrap and scaffold files into your workspace if you run the extraction script or follow install steps — run in a test workspace or dry-run first; (4) if you plan to clone the referenced GitHub repo, review that remote code before executing. Overall there are no unexplained credential requests or remote downloads, so risk is low if you review and enable intentionally.
功能分析
Type: OpenClaw Skill
Name: self-improving-negotiation
Version: 1.0.0
The skill bundle is a legitimate tool designed to help AI agents log and analyze negotiation outcomes for continuous improvement. It uses standard shell scripts (activator.sh, error-detector.sh, extract-skill.sh) and OpenClaw hooks to maintain a local directory (.learnings/) for tracking strategy failures and successes. The code contains no evidence of data exfiltration, unauthorized network access, or malicious prompt injection, and it includes explicit safety guardrails stating that the agent should never finalize agreements or approve financial terms without human intervention.
能力评估
Purpose & Capability
Name/description describe capturing negotiation learnings and promoting patterns; the scripts and hooks align with that goal (creating .learnings files, scaffolding skills, injecting reminders). No unrelated credentials, binaries, or network installs are required by the skill itself.
Instruction Scope
Runtime instructions and included scripts write/append local files (.learnings/LEARNINGS.md, NEGOTIATION_ISSUES.md, FEATURE_REQUESTS.md) and inject a virtual bootstrap reminder. The error-detector reads the environment variable CLAUDE_TOOL_OUTPUT to inspect tool output; this env var is reasonable for a PostToolUse hook but is not declared as a required env var in the registry metadata. Review that hook behavior if you run it in environments where tool outputs may contain sensitive data.
Install Mechanism
No install spec or remote downloads. All code is included in the skill bundle (bash scripts, JS/TS hook handlers). The extract-skill script scaffolds files locally; no external package registries, shorteners, or remote installers are used.
Credentials
The skill does not request credentials or config paths. It does read CLAUDE_TOOL_OUTPUT in the error-detector script (expected for a PostToolUse hook). It writes into workspace paths (project root or ~/.openclaw/workspace/.learnings) which is appropriate for its purpose but means it will create/modify files in your workspace when enabled.
Persistence & Privilege
always is false and model invocation is not disabled. Hooks are optional and must be enabled by the user; the handlers only inject a virtual reminder file and do not change other skills' configs or persist credentials. The skill does create files when scaffold/extraction scripts run, which is expected behavior for a logging/scaffolding tool.
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 self-improving-negotiation - 安装完成后，直接呼叫该 Skill 的名称或使用
/self-improving-negotiation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 of self-improving-negotiation skill.
- Enables structured logging of negotiation learnings, issues, and feature requests for continuous improvement.
- Defines clear categories for capturing strategy failures, concession leaks, weak BATNA, objection handling gaps, and agreement risks.
- Provides markdown logging templates and a quick reference guide for consistent documentation.
- Supports both OpenClaw and generic agent/CLI usage with setup and file initialization instructions.
- Promotes validated patterns to reusable assets like playbooks, objection libraries, and checklists.
元数据
常见问题
Self-Improving Negotiation 是什么？
Captures negotiation strategy failures, concession leaks, BATNA weakness, framing misses, objection handling gaps, escalation misalignment, anchor errors, an... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 83 次。
如何安装 Self-Improving Negotiation？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install self-improving-negotiation」即可一键安装，无需额外配置。
Self-Improving Negotiation 是免费的吗？
是的，Self-Improving Negotiation 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Self-Improving Negotiation 支持哪些平台？
Self-Improving 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 Self-Improving Negotiation？
由 José I. O.（@jose-compu）开发并维护，当前版本 v1.0.0。
推荐 Skills