coding-agent

Delegates coding tasks to Codex, Claude Code, Pi, or OpenCode agents via background bash processes.

Updated Mar 7, 2026
One-click install
npx skills add https://github.com/shalevamin/Tau_agent --skill coding-agent-shalevamin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: coding-agent
Source: https://github.com/shalevamin/Tau_agent/tree/main/tau-agent-main/skills/coding-agent
Command: npx skills add https://github.com/shalevamin/Tau_agent --skill coding-agent-shalevamin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Delegating substantial coding work—building features, reviewing PRs, refactoring large codebases—to CLI coding agents is error-prone: agents hang without a PTY, refuse to run outside git repos, or silently fail in the background with no visibility. This Skill provides battle-tested invocation patterns, monitoring workflows, and safety rules for running coding agents reliably. ## Core Features & Use Cases - Multi-Agent Support: Correct invocation patterns for Codex, Claude Code, OpenCode, and Pi, including when PTY mode is required and which flags to use (e.g., --full-auto, --print --permission-mode bypassPermissions). - Background Execution & Monitoring: Run agents in background sessions and track them with process actions (list, poll, log, write, submit, kill) without blocking your workflow. - Parallel Workflows: Review batches of PRs or fix multiple issues simultaneously using git worktrees, with one agent per branch. - Use Case: You need to fix issues #78 and #99 in parallel. Create two git worktrees, launch a Codex agent in each with pty:true and background:true, monitor progress via process logs, then push branches and open PRs. ## Quick Start Ask the agent to build a feature in your project, for example: "Use Codex in background mode to build a dark mode toggle in ~/Projects/myapp and notify me when it's done."

Frequently Asked Questions about coding-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run Codex CLI in the background on a coding task?▼

Run Codex with bash using pty:true, background:true, and a workdir pointing at your project, for example with codex exec --full-auto 'your task'. The call returns a sessionId you can monitor with process log and poll actions.

How do I review a pull request with a coding agent safely?▼

Clone the repository into a temporary directory or create a git worktree, check out the PR branch, then run codex review --base origin/main with pty:true in that directory. Never review PRs inside the live agent's own project folder.

What is the difference between running Claude Code and Codex from bash?▼

Claude Code should run without PTY using --print --permission-mode bypassPermissions, since PTY with --dangerously-skip-permissions can exit at the confirmation dialog. Codex, Pi, and OpenCode are interactive terminal apps and require pty:true.

Why does Codex fail when run outside a git repository?▼

Codex refuses to run outside a trusted git directory. For scratch work, create a temp directory with mktemp -d, run git init inside it, and then invoke codex exec with your prompt.

How do I fix multiple GitHub issues in parallel with coding agents?▼

Create a separate git worktree per issue, then launch one background Codex session per worktree with pty:true. Monitor all sessions with process list and log, push each fix branch, and open PRs with gh pr create.

When should I not delegate a task to a coding agent?▼

Skip delegation for simple one-line fixes you can edit directly, for reading code where a read tool suffices, and for any work inside the agent's own workspace directory, where spawning agents is forbidden.