new-loop

A command that designs and sets up an agentic loop: a program where an AI agent repeatedly works toward a goal. It requires a binary, machine-checkable definition of success and adds limits and human approval points.

In plain words
What is it for?
Use it to turn a clearly testable objective into a repeatable agent workflow with an exit check, iteration limit, spending limit, sandbox, and human checkpoint.
Why use it?
It prevents vague goals from becoming uncontrolled, expensive, or endless automated runs.

Command for Claude Code

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 commands/fltman/loop-engineer/new-loop
Clone the repo
git clone --depth 1 https://github.com/fltman/loop-engineer

Made for: Claude Code.

Per session 13 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 193 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00013 $0.00193
Opus 5 $0.00006 $0.00097
Sonnet 5 $0.00003 $0.00039
Haiku 4.5 $0.00001 $0.00019

Measured 2d ago against content hash b22aa00f589e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

new-loop 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.

.claude/commands/new-loop.md · 19 lines

What it actually says

Use the loop-architect agent to design and scaffold an agentic loop.

Goal: $ARGUMENTS

Rules:

  • If the goal is empty, ask the user what "done" looks like.
  • If the goal is not a binary, machine-checkable condition, your FIRST step is to convert it into one (or tell the user it isn't loop-ready) — do not scaffold a loop against a vague objective.
  • Pick the pattern from the loop-engineer skill's decision tree, scaffold the matching template, and wire all five guardrails (verifiable exit condition, max-iteration cap, in-code budget cap, sandbox, human checkpoint).
  • Hand off with: how to run it, how to stop it, and the per-iteration cost.
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 · 19 lines · 13 tokens per session scan A b22aa00f589e

Subscribe to this mod's changes

new-loop is a command published in the GitHub repository fltman/loop-engineer (35 stars, last pushed 2mo ago), licensed MIT. It adds 13 tokens to every session and 193 once invoked, about $0.0001 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.