Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/grubbylee/skill-manager/skill-navigatornpx skills add GrubbyLee/skill-manager --skill skill-navigatorgit clone --depth 1 https://github.com/GrubbyLee/skill-managerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/grubbylee/skill-manager/skill-navigator)<a href="https://agentmods.dev/skills/grubbylee/skill-manager/skill-navigator"><img src="https://agentmods.dev/badge/skills/grubbylee/skill-manager/skill-navigator.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00083 | $0.01879 |
| Opus 5 | $0.00042 | $0.00940 |
| Sonnet 5 | $0.00017 | $0.00376 |
| Haiku 4.5 | $0.00008 | $0.00188 |
Grade A, and why
skill-navigator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
skill-navigator:skill 清单导航
本 skill 是 skill-manager 的 AIDE 桥接入口。它只解决一个问题:当用户描述要做的事时,告诉用户本机已安装的哪一款 skill 最适合处理。
默认以 skm ask "<任务描述>" --json 为主入口,输出更口语化的首选、理由和备选;在需要更详细候选或关键词检索时,再回退到 skm recommend 和 skm search。
回答“该用哪个 skill”时,一律通过本机 skm 获取推荐结果,不要手动遍历 ~/.claude/skills、~/.codex/skills 等目录,也不要凭记忆猜测。
若 skm 命令不存在,不要猜测安装路径;提示用户运行 npm i -g aide-skill-manager 后再执行 skm setup,或在 skill-manager 项目目录重新运行 node scripts/install.mjs,并检查 npm link / PATH 是否生效。
安装与更新
推荐安装:
npm i -g aide-skill-manager
skm setup
skm scan
skm setup 会自动把本 skill 安装到 Claude Code 和 Codex 对应的用户 skill 目录。
Cursor / Gemini 手动安装
本 skill 本身是纯 Markdown 指令,不依赖平台专属能力,所以也可以在 Cursor / Gemini 的 skill 目录中手动放置一份。
复制目录:
# Cursor(示例,具体路径以本机 Cursor 配置为准)
cp -R integrations/skill-navigator ~/.cursor/skills/skill-navigator
# Gemini(示例,具体路径以本机 Gemini CLI / 配置为准)
cp -R integrations/skill-navigator ~/.gemini/skills/skill-navigator
复制后仍需要本机已安装 skm 命令,且至少运行过一次 skm scan。
源码安装:
git clone https://github.com/GrubbyLee/skill-manager.git
cd skill-manager
node scripts/install.mjs
skm scan
升级 aide-skill-manager 后,重新运行 skm setup 刷新 Claude Code / Codex 侧的本 skill;Cursor / Gemini 侧手动维护时,重新复制一次目录即可。
本 skill 的唯一真源是 GitHub 仓库:https://github.com/GrubbyLee/skill-manager/tree/main/integrations/skill-navigator。第三方 skill hub 只作为索引入口;若平台支持 GitHub 同步,应优先指向该目录,而不是维护平台侧的独立副本。
主流程
按以下顺序处理“该用哪个 skill”类问题:
-
优先使用
skm ask- 命令:
skm ask "<任务描述>" --json - 用途:直接给出首选 skill、推荐理由和备选,适合普通用户提问。
- 转述要求:保留首选名称、理由、备选;如果有
why或原因字段,直接转成自然语言,不要自己编原因。
- 命令:
-
需要更详细候选时使用
skm recommend- 命令:
skm recommend "<任务描述>" --top 5 --json - 用途:列出更多候选,便于横向比较。
- 筛选:如果用户明确只想看某个工具侧的 skill,追加
--tool <claude|codex|cursor|gemini|workbuddy|kimi>。
- 命令:
-
名称或关键词更明确时使用
skm search- 命令:
skm search "<关键词>" --json - 用途:用户已经有明确关键词,比如“markdown”“PPT”“公众号”,想快速看相关 skill。
- 命令:
-
目录可能过期时先刷新再推荐
- 触发信号:用户刚安装/删除了 skill、第一次使用、catalog 明显缺项、推荐结果里没有某个已知 skill。
- 处理方式:先告诉用户“目录可能不是最新”,并建议运行
skm scan;扫描完成后再重新执行推荐。 - 不要在未告知用户的情况下直接替用户触发
skm scan作为默认行为。
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 150 lines · 83 tokens per session scan A c642dfd74a83
skill-navigator is a skill published in the GitHub repository GrubbyLee/skill-manager (25 stars, last pushed 6d ago), licensed MIT. It adds 83 tokens to every session and 1,879 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
compose-agents-md
Meta-skill: helps create an AGENTS.md for a new project by guiding the user through selecting the right profile from the agentic library and running the compose command. Also helps create a custom AGENTS.md from scratch when no profile fits. Invoked when the user asks to set up agent instructions, create AGENTS.md, or…
deploy-config
Meta-skill: deploys the full agentic configuration to a target project — AGENTS.md composition, vendor file generation, and skill deployment — in a single workflow. Invoked when the user asks to deploy or update agent config, sync the agentic setup, or set up agents in a project.
skill-scorer
Evaluates Agent Skills (Cursor / Claude / OpenClaw compatible) and produces a quantitative, rubric-based score with actionable improvement suggestions. Use when the user asks to review, rate, audit, grade, lint, or improve a SKILL.md file, a skill folder, or a skill archive, or says things like "给这个 skill 打分", "评估一下…
vibeproof
Audit an AI-generated app for fake completeness — features that look finished in the UI but are not actually wired to anything. Traces every user-visible promise from the button to the database and back. Use when asked to run /vibeproof, to check if an app is really done, to find fake or dead UI, unwired buttons…
mariana-audit
Full-depth audit (a11y, UX, perf, SEO, security, DB, architecture, legal compliance, ops, observability) of a project. Audits by reading the source directly — installs nothing, writes nothing outside docs/audits/. Reports findings with file:line evidence and calibrated severity, optionally mitigates as discovered. Use…
rails-audit-skill
Perform comprehensive technical reviews of Ruby on Rails applications. Runs automated analysis tools (RubyCritic, Brakeman, bundler-audit, Gitleaks, Debride, linters, SimpleCov, Rails stats, Rails ERD), analyzes code for architecture, security, authorization (Pundit/CanCanCan), dead code, and design issues, and…