← 返回 Skills 市场
leegitw

Pattern Finder

作者 Lee Brown · GitHub ↗ · v1.0.2 · MIT-0
cross-platform ✓ 安全检测通过
1715
总下载
3
收藏
2
当前安装
3
版本数
在 OpenClaw 中安装
/install pattern-finder
功能描述
Discover what two sources agree on — find the signal in the noise.
安全使用建议
This skill appears coherent and low-risk, but consider: (1) It expects either raw texts or outputs from extractor skills — confirm those extractor tools are available and trustworthy. (2) The skill uses your agent's configured model: if your agent sends data to a cloud LLM, do not submit sensitive or confidential content unless you accept that exposure. (3) The skill finds agreement between sources (N=2 means 'validated' per its rubric) but agreement is not proof of truth — independently verify any high-stakes conclusions. (4) If you plan to chain this with automated workflows, test on non-sensitive examples first to ensure the extractor → comparator handoff works as you expect.
功能分析
Type: OpenClaw Skill Name: pattern-finder Version: 1.0.2 The 'pattern-finder' skill is a set of instructions for an AI agent to perform text comparison and pattern recognition between two sources. The logic is entirely contained within the SKILL.md file and focuses on identifying shared principles (N=2 validation) without external API calls, file system access, or data exfiltration. No malicious commands, obfuscation, or prompt-injection attacks were identified.
能力评估
Purpose & Capability
The skill's name/description match what the SKILL.md instructs (compare two sources and surface shared principles). One minor mismatch: the documentation expects inputs from other extractor skills ('essence-distiller' / 'pbe-extractor') but the skill metadata does not declare those as explicit dependencies — this is an interoperability note rather than a security problem.
Instruction Scope
The SKILL.md confines runtime behavior to reading two supplied sources (or their extractions), extracting principles, matching them, and returning a structured comparison. It explicitly says it does not write files and does not call third‑party services itself; it will, however, rely on whatever LLM (local or cloud) the agent is configured to use. The instructions do not ask the agent to read unrelated files, environment variables, or credentials.
Install Mechanism
No install spec and no code files — instruction-only skill. This minimizes disk/runtime install risk.
Credentials
The skill requires no environment variables, credentials, or config paths. Its declared input needs (two texts or two extractions) are proportionate to the stated purpose.
Persistence & Privilege
always is false and the skill does not request elevated persistence. Autonomous invocation is allowed by default but is not combined with other concerning privileges.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install pattern-finder
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /pattern-finder 触发
  4. 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
v1.0.2
**Pattern Finder 1.0.2 Changelog** - Updated SKILL.md to clarify no third-party APIs are called and data is analyzed strictly within your agent's configured trust boundary. - Added an explicit description of data handling and privacy, noting that this skill does not write files to disk. - Updated project homepage and tags for better discoverability. - No changes to core functionality; output format and usage remain the same.
v1.0.1
Migrated to public GitHub repo, updated homepage URLs
v1.0.0
Pattern Finder 1.0.0 initial release: - New skill for discovering shared patterns between two sources. - Identifies principles that appear in both sources, signaling N=2 validation. - Clearly separates shared, unique, and divergent principles. - Streamlined output format for easy interpretation and sharing. - Provides guidance on input requirements and when to use the skill. - Includes error messages and explains divergence analysis types.
元数据
Slug pattern-finder
版本 1.0.2
许可证 MIT-0
累计安装 2
当前安装数 2
历史版本数 3
常见问题

Pattern Finder 是什么?

Discover what two sources agree on — find the signal in the noise. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1715 次。

如何安装 Pattern Finder?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install pattern-finder」即可一键安装,无需额外配置。

Pattern Finder 是免费的吗?

是的,Pattern Finder 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Pattern Finder 支持哪些平台?

Pattern Finder 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 Pattern Finder?

由 Lee Brown(@leegitw)开发并维护,当前版本 v1.0.2。

💬 留言讨论