coding-agent

name	coding-agent
description	Run Codex CLI, Claude Code, Kiro CLI, OpenCode, or Pi Coding Agent via background process for programmatic control.
metadata	{"openclaw":{"emoji":"🧩","requires":{"anyBins":["claude","codex","opencode","pi","kiro-cli"]}}}

Coding Agent (bash-first)

Use bash (with optional background mode) for all coding agent work. Simple and effective.

⚠️ PTY Mode Required!

Coding agents (Codex, Claude Code, Kiro, Pi) are interactive terminal applications that need a pseudo-terminal (PTY) to work correctly. Without PTY, you'll get broken output, missing colors, or the agent may hang.

Always use pty:true when running coding agents:

# ✅ Correct - with PTY
bash pty:true command:"codex exec 'Your prompt'"

# ❌ Wrong - no PTY, agent may break
bash command:"codex exec 'Your prompt'"

Bash Tool Parameters

Parameter	Type	Description
command	string	The shell command to run
pty	boolean	Use for coding agents! Allocates a pseudo-terminal for interactive CLIs
workdir	string	Working directory (agent sees only this folder's context)
background	boolean	Run in background, returns sessionId for monitoring
timeout	number	Timeout in seconds (kills process on expiry)
elevated	boolean	Run on host instead of sandbox (if allowed)

Process Tool Actions (for background sessions)

Action	Description
list	List all running/recent sessions
poll	Check if session is still running
log	Get session output (with optional offset/limit)
write	Send raw data to stdin
submit	Send data + newline (like typing and pressing Enter)
send-keys	Send key tokens or hex bytes
paste	Paste text (with optional bracketed mode)
kill	Terminate the session

---

Quick Start: One-Shot Tasks

For quick prompts/chats, create a temp git repo and run:

# Quick chat (Codex needs a git repo!)
SCRATCH=$(mktemp -d) && cd $SCRATCH && git init && codex exec "Your prompt here"

# Or in a real project - with PTY!
bash pty:true workdir:~/Projects/myproject command:"codex exec 'Add error handling to the API calls'"

Why git init? Codex refuses to run outside a trusted git directory. Creating a temp repo solves this for scratch work.

---

The Pattern: workdir + background + pty

For longer tasks, use background mode with PTY:

# Start agent in target directory (with PTY!)
bash pty:true workdir:~/project background:true command:"codex exec --full-auto 'Build a snake game'"
# Returns sessionId for tracking

# Monitor progress
process action:log sessionId:XXX

# Check if done
process action:poll sessionId:XXX

# Send input (if agent asks a question)
process action:write sessionId:XXX data:"y"

# Submit with Enter (like typing "yes" and pressing Enter)
process action:submit sessionId:XXX data:"yes"

# Kill if needed
process action:kill sessionId:XXX

Why workdir matters: Agent wakes up in a focused directory, doesn't wander off reading unrelated files (like your soul.md 😅).

---

Codex CLI

Model: gpt-5.2-codex is the default (set in ~/.codex/config.toml)

Flags

Flag	Effect
exec "prompt"	One-shot execution, exits when done
--full-auto	Sandboxed but auto-approves in workspace
--yolo	NO sandbox, NO approvals (fastest, most dangerous)

Building/Creating

# Quick one-shot (auto-approves) - remember PTY!
bash pty:true workdir:~/project command:"codex exec --full-auto 'Build a dark mode toggle'"

# Background for longer work
bash pty:true workdir:~/project background:true command:"codex --yolo 'Refactor the auth module'"

Reviewing PRs

⚠️ CRITICAL: Never review PRs in OpenClaw's own project folder! Clone to temp folder or use git worktree.

# Clone to temp for safe review
REVIEW_DIR=$(mktemp -d)
git clone https://github.com/user/repo.git $REVIEW_DIR
cd $REVIEW_DIR && gh pr checkout 130
bash pty:true workdir:$REVIEW_DIR command:"codex review --base origin/main"
# Clean up after: trash $REVIEW_DIR

# Or use git worktree (keeps main intact)
git worktree add /tmp/pr-130-review pr-130-branch
bash pty:true workdir:/tmp/pr-130-review command:"codex review --base main"

Batch PR Reviews (parallel army!)

# Fetch all PR refs first
git fetch origin '+refs/pull/*/head:refs/remotes/origin/pr/*'

# Deploy the army - one Codex per PR (all with PTY!)
bash pty:true workdir:~/project background:true command:"codex exec 'Review PR #86. git diff origin/main...origin/pr/86'"
bash pty:true workdir:~/project background:true command:"codex exec 'Review PR #87. git diff origin/main...origin/pr/87'"

# Monitor all
process action:list

# Post results to GitHub
gh pr comment <PR#> --body "<review content>"

---

Claude Code

# With PTY for proper terminal output
bash pty:true workdir:~/project command:"claude 'Your task'"

# Background
bash pty:true workdir:~/project background:true command:"claude 'Your task'"

---

Kiro CLI (AWS)

AWS AI coding assistant with session persistence, custom agents, steering, and MCP integration.

Installation

# macOS / Linux (recommended)
curl -fsSL https://cli.kiro.dev/install | bash

# Linux AppImage (portable, no install needed)
wget https://desktop-release.q.us-east-1.amazonaws.com/latest/kiro-cli.appimage
chmod +x kiro-cli.appimage && ./kiro-cli.appimage

# Ubuntu (.deb)
wget https://desktop-release.q.us-east-1.amazonaws.com/latest/kiro-cli.deb
sudo dpkg -i kiro-cli.deb && sudo apt-get install -f

# Linux zip (check glibc version first: ldd --version)
# Use standard version for glibc >= 2.34, musl version for older
curl --proto '=https' --tlsv1.2 -sSf 'https://desktop-release.q.us-east-1.amazonaws.com/latest/kirocli-x86_64-linux.zip' -o 'kirocli.zip'
unzip kirocli.zip && ./kirocli/install.sh

Proxy (enterprise, v1.8.0+): Set HTTP_PROXY, HTTPS_PROXY, NO_PROXY environment variables.

Authentication

Supports GitHub, Google, AWS Builder ID, and IAM Identity Center.

kiro-cli login      # Sign in (prompts automatically on first use)
kiro-cli logout     # Sign out
kiro-cli whoami     # Check current user
kiro-cli doctor     # Diagnose issues

Remote login (SSH): Builder ID/IAM uses device code (no setup needed). For social login (Google/GitHub), use SSH port forwarding:

# 1. Run kiro-cli login on remote, note the port (e.g., 49153)
# 2. On local machine: ssh -L 49153:localhost:49153 -N user@remote-host
# 3. Press Enter on remote, complete auth in local browser

Basic Usage

kiro-cli                           # Start interactive chat (default)
kiro-cli chat "Your question"      # Direct question
kiro-cli --agent my-agent          # Use specific agent
kiro-cli chat --resume             # Resume last session (per-directory)
kiro-cli chat --resume-picker      # Pick from saved sessions
kiro-cli chat --list-sessions      # List all sessions

Non-Interactive Mode (scripting/automation)

# Single response to STDOUT, then exit
kiro-cli chat --no-interactive "Show current directory"

# Trust all tools (no confirmation prompts)
kiro-cli chat --no-interactive --trust-all-tools "Create hello.py"

# Trust specific tools only (comma-separated)
kiro-cli chat --no-interactive --trust-tools "fs_read,fs_write" "Read package.json"

🔐 Tool Trust: Use --trust-all-tools for automation (default). For untrusted input or sensitive systems, consider --trust-tools "fs_read,fs_write,shell" to limit scope.

OpenClaw Integration

# Interactive session (background)
bash pty:true workdir:~/project background:true command:"kiro-cli"

# One-shot query (non-interactive)
bash pty:true workdir:~/project command:"kiro-cli chat --no-interactive --trust-all-tools 'List all TODO comments in src/'"

# With specific agent
bash pty:true workdir:~/project background:true command:"kiro-cli --agent aws-expert 'Set up Lambda'"

# Resume previous session
bash pty:true workdir:~/project command:"kiro-cli chat --resume"

Key Arguments

Argument	Description
--no-interactive	Single response to STDOUT, then exit
--resume / -r	Resume last conversation in current directory
--resume-picker	Interactive session picker
--agent <name>	Use specified custom agent
--trust-all-tools	Skip all tool confirmations
--trust-tools <list>	Trust only listed tools (comma-separated)
-v / -vv / -vvv	Increase verbosity

In-Chat Commands

Command	Description
/editor	Open editor for multi-line input (default: vi)
/chat save <path>	Save session to JSON file
/chat load <path>	Load session from JSON file
/chat resume	Interactive session picker

Multi-line input: Press Ctrl+J to insert newline, or use /editor.

Custom Agents

Pre-define tool permissions, context resources, and behaviors:

kiro-cli agent list              # List available agents
kiro-cli agent create my-agent   # Create new agent
kiro-cli agent edit my-agent     # Edit agent config
kiro-cli agent validate ./a.json # Validate config file
kiro-cli agent set-default my-agent  # Set default

Benefits: Pre-approve trusted tools, limit tool access, auto-load project docs, share configs across team.

Steering (Project Context)

Provide persistent project knowledge via markdown files in .kiro/steering/:

.kiro/steering/
├── product.md       # Product overview
├── tech.md          # Tech stack
├── structure.md     # Project structure
└── api-standards.md # API conventions

• Workspace steering: .kiro/steering/ — applies to current project only
• Global steering: ~/.kiro/steering/ — applies to all projects
• AGENTS.md support: Place in project root or ~/.kiro/steering/

In custom agents: Add "resources": ["file://.kiro/steering/**/*.md"] to config.

MCP Integration

Connect external tools and data sources via Model Context Protocol:

kiro-cli mcp add --name my-server --command "node server.js" --scope workspace
kiro-cli mcp list [workspace|global]
kiro-cli mcp status --name my-server
kiro-cli mcp remove --name my-server --scope workspace

Other Commands

kiro-cli translate "list all files"    # Natural language → shell command
kiro-cli translate -n 3 "search text"  # Generate multiple options
kiro-cli update                        # Update Kiro CLI
kiro-cli settings list --all           # Show all settings
kiro-cli uninstall                     # Uninstall (macOS)

Plan Agent

Plan Agent is a built-in agent for structured planning before execution. It helps transform ideas into detailed implementation plans.

When to suggest Plan Agent to user:

• Complex multi-step features (e.g., "build a user authentication system")
• Requirements are unclear or need clarification
• Large features that benefit from task breakdown

When NOT to use Plan Agent:

• Simple queries or single-step tasks
• User already has clear, specific instructions
• Quick fixes or small changes

How to use:

# Start Plan mode
> /plan
# Or with immediate prompt
> /plan Build a REST API for user authentication
# Or toggle with keyboard shortcut
Shift + Tab

Plan workflow (4 phases):

1. Requirements gathering — Structured questions with multiple choice options
   • Answer with format: 1=a, 2=b or provide custom answers
2. Research and analysis — Explores codebase, identifies patterns and technologies
3. Implementation plan — Creates detailed task breakdown with clear objectives
4. Handoff to execution — After approval (y), plan transfers to execution agent

Example interaction:

> /plan I want to build a todo app

[plan] > I understand you want to build a todo app.

[1]: What platform should this todo app target?
  a. **Web Application**
  b. **Mobile App**
  c. **CLI Tool**

[2]: What's the primary use case?
  a. **Personal Task Management**
  b. **Team Collaboration**

> 1=c, 2=a

[plan] > *Researching your codebase...*
        *Creating implementation plan...*

**Implementation Plan - Todo CLI**
Task 1: Create database schema...
Task 2: Implement CLI commands...
Task 3: Add advanced features...

Does this plan look good? [y/n]: y

[default] > Implement this plan: [Plan transferred to execution agent]

Plan Agent limitations (read-only):

• ✓ Can read files, search code, research documentation
• ✗ Cannot write files or execute commands (until handoff)

---

OpenCode

bash pty:true workdir:~/project command:"opencode run 'Your task'"

---

Pi Coding Agent

# Install: npm install -g @mariozechner/pi-coding-agent
bash pty:true workdir:~/project command:"pi 'Your task'"

# Non-interactive mode (PTY still recommended)
bash pty:true command:"pi -p 'Summarize src/'"

# Different provider/model
bash pty:true command:"pi --provider openai --model gpt-4o-mini -p 'Your task'"

Note: Pi now has Anthropic prompt caching enabled (PR #584, merged Jan 2026)!

---

Parallel Issue Fixing with git worktrees

For fixing multiple issues in parallel, use git worktrees:

# 1. Create worktrees for each issue
git worktree add -b fix/issue-78 /tmp/issue-78 main
git worktree add -b fix/issue-99 /tmp/issue-99 main

# 2. Launch agent in each (background + PTY!)
bash pty:true workdir:/tmp/issue-78 background:true command:"pnpm install && codex --yolo 'Fix issue #78: <description>. Commit and push.'"
bash pty:true workdir:/tmp/issue-99 background:true command:"pnpm install && codex --yolo 'Fix issue #99: <description>. Commit and push.'"

# 3. Monitor progress
process action:list
process action:log sessionId:XXX

# 4. Create PRs after fixes
cd /tmp/issue-78 && git push -u origin fix/issue-78
gh pr create --repo user/repo --head fix/issue-78 --title "fix: ..." --body "..."

# 5. Cleanup
git worktree remove /tmp/issue-78
git worktree remove /tmp/issue-99

---

⚠️ Rules

1. Always use pty:true - coding agents need a terminal!
2. Respect tool choice - if user asks for Kiro, use Kiro; if they ask for Codex, use Codex.
   • Orchestrator mode: do NOT hand-code patches yourself.
   • If an agent fails/hangs, respawn it or ask the user for direction, but don't silently take over.
3. Be patient - don't kill sessions because they're "slow"
4. Monitor with process:log - check progress without interfering
5. --full-auto/--yolo for Codex building - auto-approves changes
6. --trust-all-tools for Kiro automation - skips confirmation prompts; use --trust-tools for untrusted input
7. --no-interactive for Kiro one-shots - single response mode
8. Parallel is OK - run many agent processes at once for batch work
9. NEVER start agents in ~/clawd/ - it'll read your soul docs and get weird ideas about the org chart!
10. NEVER checkout branches in ~/Projects/openclaw/ - that's the LIVE OpenClaw instance!
11. Suggest Kiro /plan for complex tasks - when requirements are unclear or task is multi-step, suggest Plan Agent to user and let them decide

---

Progress Updates (Critical)

When you spawn coding agents in the background, keep the user in the loop.

• Send 1 short message when you start (what's running + where).
• Then only update again when something changes:
  • a milestone completes (build finished, tests passed)
  • the agent asks a question / needs input
  • you hit an error or need user action
  • the agent finishes (include what changed + where)
• If you kill a session, immediately say you killed it and why.

This prevents the user from seeing only "Agent failed before reply" and having no idea what happened.

---

Auto-Notify on Completion

For long-running background tasks, append a wake trigger to your prompt so OpenClaw gets notified immediately when the agent finishes (instead of waiting for the next heartbeat):

... your task here.

When completely finished, run this command to notify me:
openclaw system event --text "Done: [brief summary of what was built]" --mode now

Example (Codex):

bash pty:true workdir:~/project background:true command:"codex --yolo exec 'Build a REST API for todos.

When completely finished, run: openclaw system event --text \"Done: Built todos REST API with CRUD endpoints\" --mode now'"

This triggers an immediate wake event — gets pinged in seconds, not 10 minutes.

---

Learnings (Jan 2026)

• PTY is essential: Coding agents are interactive terminal apps. Without pty:true, output breaks or agent hangs.
• Git repo required: Codex won't run outside a git directory. Use mktemp -d && git init for scratch work.
• exec is your friend: codex exec "prompt" runs and exits cleanly - perfect for one-shots.
• submit vs write: Use submit to send input + Enter, write for raw data without newline.
• Sass works: Codex responds well to playful prompts. Asked it to write a haiku about being second fiddle to a space lobster, got: "Second chair, I code / Space lobster sets the tempo / Keys glow, I follow" 🦞

---

Learnings (Feb 2026)

• Kiro CLI: AWS coding assistant with unique features — session persistence, custom agents, steering, MCP.
• No git required: Unlike Codex, Kiro runs in any directory.
• Flexible auth: GitHub/Google/AWS Builder ID/IAM Identity Center.
• Remote login: Builder ID/IAM uses device code (easy). Social login needs SSH port forwarding.
• Tool trust: --trust-all-tools for full automation, --trust-tools for fine-grained control.
• Session resume: --resume continues last conversation, --resume-picker for history browser.
• Steering: .kiro/steering/*.md provides persistent project context. Also supports AGENTS.md standard.
• Kiro + OpenClaw: --no-interactive for simple queries; interactive mode for multi-turn tasks (can run foreground or background).
• Plan Agent option: For complex multi-step tasks in Kiro, suggest /plan to user — it clarifies requirements, researches codebase, creates task breakdown, then hands off to execution.