coding-agent

Run interactive coding agents in isolated workspaces with pseudo-terminal control.

Updated May 8, 2026
One-click install
npx skills add https://github.com/freire19/Mythos --skill coding-agent-freire19
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: coding-agent
Source: https://github.com/freire19/Mythos/tree/main/skills/coding-agent
Command: npx skills add https://github.com/freire19/Mythos --skill coding-agent-freire19

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires codex, opencode, pi.

What problem does it solve?

Coding the same feature across a real codebase—especially when files must be explored, refactors are large, or PRs need review—often takes time and careful coordination that’s hard to do with simple one-shot edits.

Core Features & Use Cases

  • Runs interactive coding agents via PTY: Spawn Codex, Pi, or OpenCode reliably using pseudo-terminal mode, so terminal UI flows work correctly.
  • Focused workspace control: Provide a workdir so the agent only “sees” the intended folder, reducing accidental browsing of unrelated files.
  • Background sessions for long tasks: Start work asynchronously and monitor it with session actions (log/poll/write/submit/kill).
  • Safe PR review workflow: Review pull requests from a temp clone or isolated git worktree instead of your live folder.

Quick Start

Start an agent for a longer refactor in a dedicated directory by running: bash pty:true workdir:~/project background:true command:"codex exec --full-auto 'Refactor the authentication module and add tests'".

Frequently Asked Questions about coding-agent

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

FAQPage Schema
How do I run interactive coding agents for large code refactoring without breaking terminal UI flows?▼

You can run interactive coding agents like Codex, Pi, or OpenCode for large code refactoring by using pseudo-terminal mode (pty:true), which ensures terminal UI flows work correctly during complex file modifications.

What's the best way to review pull requests in an isolated environment using a CLI agent?▼

Review pull requests safely by running a CLI agent in a temp clone or isolated git worktree, using the workdir parameter to constrain context and prevent accidental browsing of unrelated files in your live folder.

Can I manage long-running background sessions for iterative file exploration workflows?▼

Yes, you can manage long-running background sessions for iterative file exploration by starting work asynchronously and monitoring it with session actions like list, poll, log, write, submit, and kill.

Do I need to set a specific working directory when delegating complex coding tasks to an agent?▼

Yes, you need to provide a workdir when delegating complex coding tasks so the agent only sees the intended folder, reducing accidental browsing of unrelated files and constraining the operational context.

How does pseudo-terminal handling affect interactive CLI commands for feature development?▼

Pseudo-terminal handling enables interactive CLI commands to execute reliably during feature development by spawning agents with correct PTY mode, ensuring terminal interfaces and interactive prompts function as expected.

Why should I use isolated repositories for agent-driven code refactoring and PR review?▼

Isolated repositories prevent unwanted modifications to your live folder during agent-driven code refactoring and PR review by constraining the agent's context to a dedicated workdir or temporary git worktree.