What is simmer CLAUDE.md?

Project instructions for Simmer, an iterative workflow that repeatedly improves a file, workspace, pipeline, or configuration against chosen criteria.

Use it when refining an artifact through repeated review, using human judgment, runnable checks, or both.

It provides a defined process for generating revisions, judging them, tracking the best version, and carrying feedback into the next round.

How 2389-research/simmer configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to 2389-research/simmer. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/2389-research/simmer/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/2389-research/simmer

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/2389-research/simmer/claude-md/github.svg)](https://agentmods.dev/instructions/2389-research/simmer/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/2389-research/simmer/claude-md"><img src="https://agentmods.dev/badge/instructions/2389-research/simmer/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/2389-research/simmer/claude-md"><img src="https://agentmods.dev/badge/instructions/2389-research/simmer/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 1,220 This file is loaded in full into every session.

When invoked 1,220 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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 5.1	$0.01220	$0.01220
Opus 5.5	$0.00488	$0.00488
Sonnet 5.5	$0.00244	$0.00244
Haiku 4.5	$0.00122	$0.00122

Measured 2mo ago against content hash d094a50c64f4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

simmer CLAUDE.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 2mo 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.md · 125 lines

How it starts

The opening of the file, as written. The whole thing — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Simmer Plugin

Overview

Simmer is an iterative refinement skill. It takes any artifact — a single file, a workspace of multiple files, a pipeline, a configuration — and hones it repeatedly against user-defined criteria. Evaluation can be subjective (judge-based), objective (runnable script), or hybrid (both).

Skills Included

Skill	Triggers	Description
`simmer`	"simmer this", "refine this", "hone this", "iterate on this", "optimize this pipeline"	Orchestrator: setup → loop → output

Subskills (invoked by orchestrator)

Subskill	Role
`simmer:simmer-setup`	Identify artifact (file or workspace), elicit criteria, determine evaluation method, produce setup brief
`simmer:simmer-generator`	Produce improved candidate from current + ASI feedback + background context
`simmer:simmer-judge`	Score candidate 1-10 per criterion (with optional evaluator output), produce ASI
`simmer:simmer-judge-board`	Multi-judge panel with deliberation — drop-in replacement for simmer-judge when JUDGE_MODE is board
`simmer:simmer-reflect`	Record trajectory, track best-so-far, pass ASI forward

Flow

"Simmer this" / "Refine this" / "Optimize this pipeline"
    ↓
Setup (identify artifact + criteria + evaluation method)
    ↓
Iteration 0: Evaluate the seed
    ↓
Loop N times:
  Generate (address ASI) → Evaluate (run evaluator if present) → Judge (score + new ASI) → Reflect (record)
    ↓
Output best candidate + trajectory

Key Design Decisions

Context discipline: Generator doesn't see scores (avoids optimizing for numbers). Judge doesn't see intermediate scores (avoids anchoring). Reflect is the only subskill that sees the full trajectory.

Seed calibration: Judge receives the seed artifact + iteration-0 scores on every iteration as a fixed calibration reference. Can score above or below the seed. This compresses score variance across runs.

ASI (Actionable Side Information): Judge identifies the highest-leverage direction each round. For single-file targets, this is a single focused fix. For workspace targets, this is a single strategic direction that may involve coordinated changes across multiple files. Focused directional improvement compounds better than scattered edits.

Read the full file on GitHub · 125 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.

2mo ago First seen · 125 lines · 1,220 tokens per session scan A d094a50c64f4

Subscribe to this mod's changes

simmer CLAUDE.md is an instructions file published in the GitHub repository 2389-research/simmer (16 stars, last pushed 5d ago), licensed MIT. It adds 1,220 tokens to every session, about $0.0049 per session on Opus 5.5. 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.

artifact criteria evaluator generator improvement iteration

simmer CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
simmer CLAUDE.md2389-research/simmer	Instructions file	16	5d ago	MIT	A	1,220	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

simmer: Instructions file for Claude Code