What is loop AGENTS.md?

A set of instructions for developing Loop, a command-line tool that coordinates Claude and Codex coding sessions.

Use it when changing Loop’s planning, execution, dashboard, tmux, worktree, resume, testing, build, or release behavior.

It records the commands and workflows needed to preserve interactive runs, saved session state, dashboards, terminal workspaces, and separate working folders.

loop AGENTS.md — Instructions file by axeldelafosse

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/axeldelafosse/loop/agents-md

Clone the repo

git clone --depth 1 https://github.com/axeldelafosse/loop

Made for: Codex, OpenCode.

Per session 665 This file is loaded in full into every session.

When invoked 665 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.

Model	Per session	Once invoked
Fable 5	$0.00665	$0.00665
Opus 5	$0.00332	$0.00332
Sonnet 5	$0.00133	$0.00133
Haiku 4.5	$0.00067	$0.00067

Measured 2d ago against content hash d3ac41c45b52, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

loop 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

Proofloop AGENTS.md — 100% identical, 0 lines differ

AGENTS.md · 53 lines

How it starts

The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Please keep the code dead-simple and keep the src/loop/main.ts file under 150 lines of code.

Quick Commands

Format code: bun run fix
Check lint/types/style: bun run check
Run tests: bun test
Build executable: bun run build
Install global binary/aliases: bun run install:global
Cut a patch release: bun run release:patch

Repo Workflows

Plain-text prompts auto-create PLAN.md first, then optionally run --review-plan; if you are changing planning behavior, keep that flow aligned.
Default CLI behavior is paired Claude/Codex execution with persisted run state under ~/.loop/runs; preserve --run-id / --session resume behavior when changing startup, planning, or loop flow.
Running loop with no args starts the paired interactive tmux workspace (--tmux); loop dashboard opens the live panel for active sessions, loop-owned paired runs, and tmux sessions. Keep panel-only changes separate from task-running changes when possible.
--tmux and --worktree are first-class execution modes. In paired mode, tmux opens Claude/Codex side-by-side and resumed run ids should stay aligned with matching tmux/worktree names.
--claude-only and --codex-only switch out of the default paired flow; keep single-agent behavior working when changing shared CLI parsing or resume logic.
loop update / loop upgrade are supported manual update commands for installed binaries; source runs should continue to rely on git pull.
When options are provided without a prompt, loop reuses PLAN.md if it already exists; keep that fallback aligned with the plain-text prompt planning flow.

Coding Standards

Keep functions small and easy to read.
Use explicit parameter/return types when they improve clarity.
Prefer unknown over any for unknown values.
Use const by default, let only when reassignment is needed.
Prefer for...of, optional chaining, nullish coalescing, and template literals.
Use early returns to reduce nesting.
Use async/await instead of promise chains.
Throw Error objects with clear messages.
Extract magic numbers/strings into named constants when reused.

Read the full file on GitHub · 53 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.

2d ago First seen · 53 lines · 665 tokens per session scan A d3ac41c45b52

Subscribe to this mod's changes

loop AGENTS.md is an instructions file published in the GitHub repository axeldelafosse/loop (277 stars, last pushed 5mo ago), licensed MIT. It adds 665 tokens to every session, about $0.0033 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.

Analytics and metrics Git and workflow Testing bun claude claude-code claude-code-channels codex codex-cli

loop AGENTS.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

Copies of this mod

How it starts

Quick Commands

Repo Workflows

Coding Standards

What this file has done since we first saw it

Other instructions, from other repositories

codex-studio AGENTS.md

agentpulse AGENTS.md

deepwiki-by-cc CLAUDE.md

session-map AGENTS.md

agentpulse CLAUDE.md

agent-thread AGENTS.md