new-collection

Instructions for creating a collection: a named, versioned bundle of issue types that describes a software-development process for AI agents.

In plain words
What is it for?
Use them to design and author shareable AI workflows with ordered steps, review gates, and rejection paths.
Why use it?
They clarify the difference between a collection, an issue type, and an Operator workflow so the new workflow has the right structure.

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/untra/operator/new-collection
Clone the repo
git clone --depth 1 https://github.com/untra/operator

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 2,818 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 $0.02818
Opus 5 $0.00006 $0.01409
Sonnet 5 $0.00003 $0.00564
Haiku 4.5 $0.00001 $0.00282

Measured 2d ago against content hash d6be6cda2a5c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

new-collection 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-collection.md · 265 lines

How it starts

The opening of the file, as written. The whole thing — 265 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Author a New Collection

Create a new workflow collection: a named, versioned bundle of issue types that encodes a coherent process of doing software development with AI agents.

The mechanics are easy; the design is the hard part. workflows allow agents to apply themselves in a deterministic manner.

Vocabulary — get this right first

Three terms that are easy to conflate:

Term What it is
Operator workflow The step graph itself: an ordered set of typed steps with review gates and reject edges. Lives in an issue type's steps. This is the native format.
Issue type One kind of work (FEAT, PRD, ELVSTAGE). Carries identity, input fields, and exactly one Operator workflow.
Collection A bundle of issue types that work together. What you are authoring.

Kanban issue types describe how a team labels work; a collection describes how the agents do the work.

Other non-Operator "workflows"

the term "Workflow" is overloaded in the AI space. Operator workflows are json described collections of issuetypes. they can export to other workflows, such as claude workflows among others. As a result Operator workflows are meant to compose as broad and neutral as possible, with some opinions about structure and behavior. Define your workflows in the context of the work you want to get done.

Step 1 — Decide what loop you are encoding

Answer these before opening an editor. If you cannot answer them crisply, the collection is not ready to write.

  1. What is the loop? What repeats, and what ends it? "Implement one right-sized story per fresh agent context" (ralph_loop) is a loop. "Do software development" is not.
  2. What are the work types? Each issue type must be a genuinely different shape of work, not a different priority or label. If two types have the same steps, they are one type with a field.
  3. Where does state live between steps? Agents lose context. Name the files or surfaces that carry memory forward (workflow_hints.memory_surfaces).
  4. What are the gates? Where must a human or a test approve before continuing, and what happens on rejection?
  5. When does it stop? Both success and give-up conditions.

Read the full file on GitHub · 265 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 · 265 lines · 13 tokens per session scan A d6be6cda2a5c

Subscribe to this mod's changes

new-collection is a command published in the GitHub repository untra/operator (37 stars, last pushed 5d ago), licensed MIT. It adds 13 tokens to every session and 2,818 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.