← 返回 Skills 市场
ciklopentan

skill-creator-canonical

作者 ciklopentan · GitHub ↗ · v6.1.2 · MIT-0
cross-platform ✓ 安全检测通过
245
总下载
0
收藏
0
当前安装
5
版本数
在 OpenClaw 中安装
/install skill-creator-canonical
功能描述
Create, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch, restructuring an existing skill, auditing skill quality, or making a...
安全使用建议
This appears to be a legitimate skill-authoring tool. Before running any scripts: (1) review the included Python scripts (init_skill.py, package_skill.py, quick_validate.py, etc.) to confirm behavior and any paths they will modify; (2) run these tools in an isolated or test directory (not your home or production data) because they create and write files and can package directories; (3) do not supply credentials or secret-containing paths when asked for output or skill directories; (4) if you plan to publish or run packaged output in a shared environment, confirm the .clawhubignore and packaging rules exclude any sensitive files. Overall the package is coherent with its stated purpose.
功能分析
Type: OpenClaw Skill Name: skill-creator-canonical Version: 6.1.2 The skill bundle is a legitimate developer tool designed to create, validate, and package OpenClaw AgentSkills. The included Python scripts (such as init_skill.py and package_skill.py) perform standard file and directory operations consistent with the stated purpose. Notably, package_skill.py implements security best practices by explicitly skipping symlinks and validating that files remain within the skill root to prevent path traversal or accidental data exfiltration. The SKILL.md instructions provide clear, structured guidance for the AI agent without any evidence of malicious prompt injection or unauthorized behavior.
能力标签
cryptocan-make-purchases
能力评估
Purpose & Capability
Name/description (authoring and auditing skills) align with the included files and runtime instructions: SKILL.md, authoring references, and scripts for initializing, validating, and packaging skills. Nothing requested (no env vars, no external binaries) is unexpected for this purpose.
Instruction Scope
SKILL.md and references explicitly guide filesystem operations (creating/normalizing skill folders, writing SKILL.md, adding scripts/references/assets) and call included scripts (init_skill.py, package_skill.py, validate_weak_models.py, quick_validate.py). Those actions are appropriate for an authoring tool. The instructions do not demand unrelated secrets or network exfiltration and they constrain reading additional references only when explicitly requested.
Install Mechanism
No install spec — instruction-only with bundled helper scripts. The included scripts are plain Python files; there are test scripts but no downloads, URL-based installers, or archive extraction steps from untrusted hosts.
Credentials
The skill requires no environment variables, no credentials, and no config paths. The scripts operate on user-provided filesystem paths (creating/packing skill folders), which is proportional to an authoring/packaging tool.
Persistence & Privilege
The skill is not always-enabled and does not request elevated or persistent platform privileges. It writes/reads files within skill directories (expected for authoring/packaging). Packaging code includes sane checks (e.g., skipping symlinks, checking path escapes).
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 skill-creator-canonical
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /skill-creator-canonical 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
v6.1.2
Hardened review, validation, packaging, and publish flows; aligned scaffold and archive hygiene.
v6.1.1
Tighten publish-readiness gate with PACKAGING_CHECKLIST, .clawhubignore, mandatory _meta.json, and log-file hygiene after 6-round dual-thinking review.
v6.2.0
6-round dual-thinking review; tightened path selection, review gating, blocked-validation handling, fallback triggers, and weak-model execution outputs.
v6.1.0
6-round dual-thinking review: execution-first SKILL.md, reference split, path fixes, compliance note, stronger weak-model determinism
v6.0.0
Canonical skill-authoring workflow rewrite; weak-model-first guidance and validator; dual-thinking review integrated with conditional deep-review triggers.
元数据
Slug skill-creator-canonical
版本 6.1.2
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 5
常见问题

skill-creator-canonical 是什么?

Create, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch, restructuring an existing skill, auditing skill quality, or making a... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 245 次。

如何安装 skill-creator-canonical?

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

skill-creator-canonical 是免费的吗?

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

skill-creator-canonical 支持哪些平台?

skill-creator-canonical 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 skill-creator-canonical?

由 ciklopentan(@ciklopentan)开发并维护,当前版本 v6.1.2。

💬 留言讨论