atria AGENTS.md

Project-specific coding conventions for Atria, a Go terminal interface for managing multiple AI coding agents in terminal tabs or panes. They cover code style, error handling, commit messages, and the project layout.

In plain words
What is it for?
Use them when editing Atria to follow its Go conventions, format code, return contextual errors, organize files, and write consistent commits.
Why use it?
They help changes fit Atria's existing Go practices and prevent common maintenance problems such as inconsistent formatting or unsafe error handling.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/sethdeckard/atria/agents-md
Clone the repo
git clone --depth 1 https://github.com/sethdeckard/atria

Made for: Codex, OpenCode.

Per session 4,703 This file is loaded in full into every session.
When invoked 4,703 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.04703 $0.04703
Opus 5 $0.02351 $0.02351
Sonnet 5 $0.00941 $0.00941
Haiku 4.5 $0.00470 $0.00470

Measured 2d ago against content hash 3397bd249963, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

atria AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 326 lines

How it starts

The opening of the file, as written. The whole thing — 326 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Atria - Agent Conventions

Internal development conventions. See README.md for user documentation.

Project Overview

Atria is a Go TUI tool for managing multiple AI coding agents (Claude Code, Codex, OpenCode) running in terminal tabs/panes. Built with Bubble Tea (bubbletea, lipgloss, bubbles).

Code Style

  • Idiomatic Go: follow Effective Go patterns, standard naming conventions
  • Format all code with gofmt / goimports
  • Error handling: return errors, don't panic. Wrap errors with context using fmt.Errorf("context: %w", err)
  • No log.Fatal in library code (internal packages). Only main.go may exit the process

Commit Messages

  • Subject: max 50 chars, capitalized, imperative mood (e.g. "Add feature" not "Added feature")
  • Blank line between subject and body
  • Body: wrapped at 72 chars, describe what and why, not how

Project Structure

main.go                          # Entry point
internal/
  config/config.go               # TOML config parsing
  model/types.go                 # Domain types (Project, AgentSession, etc.)
  model/store.go                 # In-memory store + JSON persistence
  terminal/backend.go            # Backend interface
  terminal/detect.go             # Agent detection from session name
  terminal/monitor.go            # Log reading + status classification
  terminal/composite.go          # Composite backend (PTY + integrations)
  terminal/cache.go              # Cached session list with TTL
  terminal/cwd.go                # CWD discovery strategies (shared)
  terminal/kitty/client.go       # Kitty backend (kitten @ CLI via socket)
  terminal/wezterm/client.go     # WezTerm backend (wezterm cli)
  terminal/iterm/proto/api.proto  # iTerm2 protobuf spec (trimmed)
  terminal/iterm/proto/api.pb.go  # Generated protobuf Go code
  terminal/iterm/conn.go         # iTerm2 WebSocket connection manager
  terminal/iterm/client.go       # iTerm2 backend (native protobuf API)
  terminal/tmux/client.go        # tmux backend
  terminal/pty/client.go         # PTY backend (built-in multiplexer)
  terminal/pty/session.go        # PTY session (process + vt10x emulator)
  tui/app.go                     # Root Bubble Tea model
  tui/messages.go                # tea.Msg types
  tui/commands.go                # tea.Cmd factories
  tui/keys.go                    # Key bindings
  tui/styles.go                  # Lip Gloss styles
  tui/settings.go                # Settings screen (config editor)
  tui/projectlist.go             # Agent list view (agents dashboard)
  tui/chatview.go                # Chat/send view
  tui/termview.go                # Embedded terminal view (PTY backend)
  tui/paths.go                   # Path display utilities (contractHome)
  tui/gitinfo.go                 # Git worktree detection

Read the full file on GitHub · 326 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 326 lines · 4,703 tokens per session scan A 3397bd249963

Subscribe to this mod's changes

atria AGENTS.md is an instructions file published in the GitHub repository sethdeckard/atria (20 stars, last pushed 1mo ago), licensed MIT. It adds 4,703 tokens to every session, about $0.0235 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.