open-code-review: Instructions file for Claude Code

CLAUDE.md

open-code-review CLAUDE.md is an instructions file for Claude Code from spencermarx/open-code-review. It costs 1,916 tokens per session, scanned A, original, Apache-2.0.

A set of project instructions for an AI coding assistant working on the open-code-review repository. It explains when to read OpenSpec proposals, how to follow TypeScript conventions, and how releases are handled through GitHub and npm.

In plain words
What is it for?
It is for guiding feature work, architecture changes, code conventions, agent-file updates, and package releases.
Why use it?
It keeps code changes, planning documents, and releases consistent with the repository's rules.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions .claude-plugin; mentions AGENTS.md.

This is spencermarx/open-code-review's own configuration. It tells Claude Code how to work on open-code-review itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything open-code-review configures →

Reuse

Borrowing it

Nothing to install: this file belongs to spencermarx/open-code-review. 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/spencermarx/open-code-review/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/spencermarx/open-code-review

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.

agentmods badge for open-code-review CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/spencermarx/open-code-review/claude-md/github.svg)](https://agentmods.dev/instructions/spencermarx/open-code-review/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/spencermarx/open-code-review/claude-md"><img src="https://agentmods.dev/badge/instructions/spencermarx/open-code-review/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.

agentmods 80×15 button for open-code-review CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/spencermarx/open-code-review/claude-md"><img src="https://agentmods.dev/badge/instructions/spencermarx/open-code-review/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 1,916 This file is loaded in full into every session.
When invoked 1,916 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.01916 $0.01916
Opus 5.5 $0.00766 $0.00766
Sonnet 5.5 $0.00383 $0.00383
Haiku 4.5 $0.00192 $0.00192

Measured 1mo ago against content hash 800e783b5b85, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

open-code-review 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 1mo 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 · 72 lines

How it starts

The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.

OpenSpec Instructions

These instructions are for AI assistants working in this project.

Always open @/openspec/AGENTS.md when the request:

  • Mentions planning or proposals (words like proposal, spec, change, plan)
  • Introduces new capabilities, breaking changes, architecture shifts, or big performance/security work
  • Sounds ambiguous and you need the authoritative spec before coding

Use @/openspec/AGENTS.md to learn:

  • How to create and apply change proposals
  • Spec format and conventions
  • Project structure and guidelines

Keep this managed block so 'openspec update' can refresh the instructions.

Code Conventions

  • TypeScript only: Do not create raw .js or .mjs files unless they serve a config purpose (e.g., vite.config.mjs, eslint.config.mjs). All project code, scripts, and utilities must be written in TypeScript.
  • Nx-native automation: Release process automation must use Nx extension points (e.g., VersionActions, preVersionCommand), not npm lifecycle scripts or standalone scripts.
  • Agent assets — edit source, then sync: Agent docs, skills, commands, references, and other agent-related files have their source of truth in packages/agents/. ALWAYS edit them there, then run nx run cli:update to write the changes out to the local project's .ocr/ directory. Never hand-edit the generated .ocr/ copies directly — they will be overwritten on the next sync and your edits will drift from source.
  • Shared layers live in packages/shared/*, apps never depend on apps (CI-enforced): cli and dashboard are application packages and MUST NOT depend on one another. This DAG is enforced by @nx/enforce-module-boundaries (root eslint.config.mjs, keyed off each project's scope:* tag) and gated in CI via nx run-many -t lint — an accidental dashboard → cli (or shared → app) import fails the build, not just review. Code both apps need (persistence, domain/state, config, cross-platform utilities) lives in dedicated library packages under packages/shared/* that each app depends on directly. The current shared packages are @open-code-review/platform (cross-platform/runtime utilities), @open-code-review/persistence (the node:sqlite adapter db + workflow state lifecycle + test-support + vendor-resume + the node:sqlite runtime precondition runtime-checks — kept in one package because db and state have a mutually-recursive type cycle (db/types.ts ↔ state/types.ts); a package boundary between them would form a dependency cycle. The single-module-instance connection-cache singleton is a consequence of that co-location, not its root cause), and @open-code-review/config (runtime-config + team-config + models).
  • Shared packages are source-only, private, and inlined — never published: each packages/shared/* package mirrors platform exactly — private: true, version 0.0.0, every exports condition (types/source/default) points at ./src/*.ts (no build.mjs, no dist), and it is declared by its consumers as a devDependency: workspace:* (consumer-side rule). A shared package still declares its own runtime third-party deps in its dependencies — they are inlined into the consumer's bundle, so they must resolve at build time. esbuild inlines the .ts source into each app's published bundle, so these packages are excluded from the release set (!packages/shared/* in nx.json) and do not join the fixed cli+agents release group. Do NOT give a shared package a build target or a dist — that machinery was removed in the cutover and must not return.
  • Graduation is by cause, not by count: a slice graduates from an app package into a packages/shared/* package the moment it is consumed across a package boundary (by the other app, an e2e package, or another shared package) rather than only by its owning app's own code. There is no subpath-count trigger. A genuinely app-internal module stays in its app; the goal is to keep the dependency graph a DAG of app → shared → shared, never app → app.

Read the full file on GitHub · 72 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. 1mo ago First seen · 72 lines · 1,916 tokens per session scan A 800e783b5b85

Subscribe to this mod's changes

open-code-review CLAUDE.md is an instructions file published in the GitHub repository spencermarx/open-code-review (382 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 1,916 tokens to every session, about $0.0077 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 2026-09-03.

Related

Other instructions, from other repositories