Clojure REPL Evaluation
技能 警告 活跃Evaluate Clojure code via nREPL using clj-nrepl-eval. Use this when you need to test code, check if edited files compile, verify function behavior, or interact with a running REPL session.
Evaluate Clojure code interactively against a running nREPL server for testing, debugging, and verifying code changes.
功能
- Discover and connect to nREPL servers
- Evaluate arbitrary Clojure code snippets
- Maintain persistent REPL sessions
- Reset REPL session state
- Handle code via arguments, heredoc, or stdin
使用场景
- Verify that edited Clojure files compile and load correctly
- Test function behavior interactively
- Check the current state of the REPL
- Debug code by evaluating expressions
非目标
- Replacing a full IDE's debugging capabilities
- Managing Clojure project dependencies directly
- Executing arbitrary shell commands outside of nREPL interaction
工作流
- Discover available nREPL servers
- Select an nREPL server (or start a new one)
- Provide Clojure code for evaluation
- Receive and review the evaluation results
先决条件
- A running nREPL server accessible via host:port
- Clojure environment set up
License
- critical:License usabilityThe bundled LICENSE-AGPL.txt is the GNU Affero General Public License v3, which is a strong copyleft license that may not be compatible with all downstream use cases and requires specific considerations for network-facing services.
Security
- warning:InjectionWhile the tool evaluates Clojure code, there's a risk of injection if untrusted Clojure code is passed without sanitization, although the SKILL.md guides users on safe code passing. No explicit protection against arbitrary code execution is detailed.
- warning:Sandbox IsolationThe skill executes arbitrary Clojure code which could potentially interact with the file system or network, and the SKILL.md does not explicitly detail sandboxing measures beyond standard nREPL server isolation.
Trust
- warning:Issues AttentionThe repository has 418 open issues and 321 closed issues in the last 90 days, indicating a high volume of activity but also a potentially slow response rate for new issues given the ratio.
Execution
- warning:ValidationThe SKILL.md mentions automatic delimiter repair for parentheses, but there's no explicit mention of a schema library for validating the structure or content of arbitrary Clojure code passed as input.
Code Execution
- warning:Error HandlingWhile the `clj-nrepl-eval` tool likely handles errors from Clojure code execution, the SKILL.md does not detail how these errors are translated into structured agent-understandable error messages or if they provide specific recovery steps.
Compliance
- info:GDPRThe skill evaluates Clojure code, which could potentially process personal data if the user's code does. There's no explicit mention of sanitization beyond what the user's code provides.
Documentation
- info:READMEThe README.md provides a high-level overview of Metabase, not the clj-nrepl-eval skill. The SKILL.md serves as the primary documentation for this skill.
Errors
- warning:Actionable error messagesThe SKILL.md outlines error handling for REPL interaction but does not detail specific actionable error messages for various failure modes, nor does it provide remediation steps or documentation links for user-facing errors.
Protocol
- warning:Idempotent retry & timeoutsThe tool includes a timeout option, but the idempotency of mutations (if any are performed by the evaluated Clojure code) is not guaranteed by the tool itself, and error handling for timeouts is not detailed in terms of structured retryable errors.
安装
npx skills add metabase/metabase通过 npx 运行 Vercel skills CLI(skills.sh)— 需要本地安装 Node.js,以及至少一个兼容 skills 的智能体(Claude Code、Cursor、Codex 等)。前提是仓库遵循 agentskills.io 格式。
质量评分
警告类似扩展
Remember Interactive Programming
100A micro-prompt that reminds the agent that it is an interactive programmer. Works great in Clojure when Copilot has access to the REPL (probably via Backseat Driver). Will work with any system that has a live REPL that the agent can use. Adapt the prompt with any specific reminders in your workflow and/or workspace.
Openclaw Debugging
100Debug OpenClaw model, provider, tool-surface, code-mode, streaming, and live/Crabbox behavior by choosing the right logs, probes, and proof path before changing code.
Agent Coder
99Agent skill for coder - invoke with $agent-coder
Clojure Write
75Guide Clojure and ClojureScript development using REPL-driven workflow, coding conventions, and best practices. Use when writing, developing, or refactoring Clojure/ClojureScript code.
使用我的浏览器
100当工作依赖于用户实时浏览器会话或可见的渲染状态，而不是静态抓取时使用，尤其是在浏览器调试上下文或 DevTools 选择的元素或请求、登录仪表板或 CMS 流程、localhost 应用、上传、下载、媒体检查、DOM 或 iframe 检查、Shadow DOM，或看起来像软 404、身份验证墙、反机器人检查或速率限制的浏览器故障时。
Node Connect
100Diagnose OpenClaw Android, iOS, or macOS node pairing, QR/setup code, route, auth, and connection failures.