duet CLAUDE.md

duet CLAUDE.md is an instructions file for coding agents from tslateman/duet. It costs 1,133 tokens per session, scanned B, original, MIT.

A set of Claude Code instructions, commands, and skills for reflection, code quality, writing, design, debugging, and project documentation.

In plain words
What is it for?
Use it to capture learnings, audit memory files, debug changes, create architecture decision records, design systems, and make diagrams.
Why use it?
It gives the coding agent repeatable ways to inspect work, record decisions, find problems, and keep project guidance organized.

Instructions file

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/tslateman/duet/claude-md
Clone the repo
git clone --depth 1 https://github.com/tslateman/duet

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.

agentmods badge for duet CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/tslateman/duet/claude-md.svg)](https://agentmods.dev/instructions/tslateman/duet/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/tslateman/duet/claude-md"><img src="https://agentmods.dev/badge/instructions/tslateman/duet/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 1,133 This file is loaded in full into every session.
When invoked 1,133 The same file — it is already loaded in full.
Security scan B 1 finding. 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 $0.01133 $0.01133
Opus 5 $0.00566 $0.00566
Sonnet 5 $0.00227 $0.00227
Haiku 4.5 $0.00113 $0.00113

Measured 3d ago against content hash 2ca393349fcf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

duet CLAUDE.md scanned grade B with 1 finding 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 3d 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.

Unrestricted tool accessmediumExcessive agency

A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.

- `allowed-tools`: Optional. Omit to allow all tools
CLAUDE.md · 122 lines

How it starts

The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Duet

Claude Code plugin for reflection, code quality, writing, and design. Markdown only, no build system.

Lint

markdownlint commands/**/*.md skills/**/SKILL.md

Project Structure

commands/
  capture-memory.md    Capture learnings to MEMORY.md
  lint.md              Run markdownlint on files
  memory-audit.md      Audit project memory files for staleness
  ponder.md            Surface a question worth sitting with
  probe.md             Probe recent changes with pointed questions
skills/
  FRAMEWORKS.md        Framework-to-skill index
  adr/                 Architecture Decision Records (Nygard)
  automagic-problem-discovery/   Friction audit that ends in a built tool (5 Whys)
  caveman/                       Ultra-compressed output mode (~75% token reduction)
  debrief/                       Review agent work, surface loose ends
  debugging/                     Systematic debugging (Agans)
  design/                        Intentional design (visual, API, system)
  excalidraw/                    Programmatic Excalidraw diagrams
  grill-with-docs/               Interview-driven design with CONTEXT.md/ADR upkeep
  ia/                            Information architecture (Rosenfeld + Diataxis)
  improve-codebase-architecture/ Find deepening opportunities (Ousterhout + Feathers)
  mermaid/                       Mermaid.js diagrams for GitHub markdown
  naming/                        Naming as design diagnostic (Benner + Bloch)
  performance/                   Profiling and optimization (Knuth)
  prose/                         Strunk's Elements of Style for all writing
  research/                      Systematic technical investigation
  retro/                         Session retrospective
  review/                        Structured code review with context capture
  sweep/                         Post-op audit for artifacts, damage, stale refs
  testing/                       Test strategy (Beck's Test Desiderata)
  vamp/                          Survey the project, choose what to play next
  vibe-check/                    Judgment linter for vibe-coded output
  visual-recap/                  Structured visual recap of a diff (BuilderIO)
  writing-great-skills/          Reference for writing skills well (mattpocock)
  zoom-out/                      Map callers and neighbors before reading details
agents/
  code-simplifier.md   Refine code for clarity and maintainability
statusline/
  statusline-command.sh  Lean status bar (dir, git, model, context, vim)
scripts/
  generate-readme.sh   Auto-generate README tables from source
evals/
  evals.json           Master eval registry (generated from the scenario tree)
  run.sh               Eval runner (with/without skill + grader)
  <skill>/<scenario>/  43 scenarios covering 14 of the 25 skills; vibe-check has 4, the rest 3 each

Read the full file on GitHub · 122 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. 3d ago First seen · 122 lines · 1,133 tokens per session scan B 2ca393349fcf

Subscribe to this mod's changes

duet CLAUDE.md is an instructions file published in the GitHub repository tslateman/duet (1 stars, last pushed 7d ago), licensed MIT. It adds 1,133 tokens to every session, about $0.0057 per session on Opus 5. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.