Coding Agent

An AI coding agent that works where your code lives, including remote servers over SSH.

 ◆ Let me read the file first.

   ⚙ Tool  read   path=server.go
      ✓ Done: 312 lines read

 ◆ Found it — the goroutine in handleConnection() is never joined.
   I'll patch it now.

   ⚙ Tool  edit   path=server.go

   ╭─────────────────────────────────────────────────────╮
   │  - go handle(conn)                                  │
   │  + wg.Add(1)                                        │
   │  + go func() { defer wg.Done(); handle(conn) }()   │
   ╰─────────────────────────────────────────────────────╯
      ✓ Edit applied

──────────────────────────────────────────────────────────
 > _
──────────────────────────────────────────────────────────
  Model: openai / gpt-4o    Approve: Ask │ Tokens: 2104 / 128000 (2%)

---

Install

go install github.com/cnjack/jcode/cmd/jcode@latest

---

What it does

Describe a task with human language. The agent reads your files, writes changes, runs commands, and shows you exactly what it did — step by step.

• Read, edit & write files — surgical string-level edits with inline diffs
• Run shell commands — output shown in a bordered box, last N lines surfaced
• Search codebases — regex grep with ripgrep/grep fallback
• Track tasks — a live todo list the agent updates as it works through multi-step jobs
• Resume sessions — every conversation is recorded; pick up exactly where you left off

---

SSH — work on any machine

Type /ssh user@host and every tool — file reads, edits, shell commands — runs transparently on the remote host. No separate setup, no changing your workflow.

 You › /ssh deploy@10.0.1.5:/var/www/app

   ✓ SSH  Connected · linux/amd64

 You › why is nginx restarting?

 ◆ Let me check the container logs.

   ⚙ Tool  execute  [deploy@10.0.1.5]  docker logs app-nginx-1 --tail 20

   ╭─────────────────────────────────────────────────────╮
   │  nginx: [emerg] bind() to 0.0.0.0:80 failed        │
   │  (98: Address already in use)                       │
   ╰─────────────────────────────────────────────────────╯

 ◆ Port 80 is already taken. Let me find what's holding it.

Save connections as named aliases and jump between hosts with a keypress:

  ┌─────── /ssh ────────────────────────────────────┐
  │  > 🔗 prod        deploy@10.0.1.5:/var/www/app  │
  │    🔗 staging      ci@10.0.1.8:/srv/staging      │
  │    ➕ Connect New SSH                             │
  └─────────────────────────────────────────────────┘

---

More

• Any OpenAI-compatible provider — switch models mid-session with /model
• MCP servers — connect HTTP / SSE / stdio MCP servers; their tools merge with the built-ins
• Approval mode — default is Ask (confirm each tool call); toggle to Auto for unattended runs
• Session resume — /resume brings back the full conversation history for any past project session
• Plan mode — agent explores the codebase read-only and presents a plan before making changes
• Skills — built-in skills (PR review, security review, etc.) loaded on demand for domain-specific tasks
• Subagents — delegate subtasks to independent child agents for parallel research or exploration
• Background tasks — long-running commands (builds, tests) run in the background; check status anytime
• Context awareness — auto-detects Git branch, project type, and directory structure at startup

  ┌──────────────── Resume Session ─────────────────┐
  │  > 2026-03-12  gpt-4o      fix nginx crash       │
  │    2026-03-11  gpt-4o      refactor auth module  │
  │    2026-03-10  claude-3.5  add pagination logic  │
  └─────────────────────────────────────────────────┘

---

Config is created on first launch at ~/.jcoding/config.json. Run coding --doctor to verify connectivity.

---

MIT