reflex CLAUDE.md

reflex CLAUDE.md is an instructions file for coding agents from reflex-search/reflex. It costs 6,019 tokens per session, scanned A, original, MIT.

Project instructions for Reflex, a local Rust code-search engine designed for coding agents. It searches an entire codebase for text and patterns, including code uses and comments, and returns file paths, line numbers, and context.

In plain words
What is it for?
Use it to guide work on the Reflex project and explain its code-search architecture and components.
Why use it?
It gives an agent a shared understanding of the project's purpose, design, and operating principles before it searches or changes code.

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

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 reflex CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/reflex-search/reflex/claude-md.svg)](https://agentmods.dev/instructions/reflex-search/reflex/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/reflex-search/reflex/claude-md"><img src="https://agentmods.dev/badge/instructions/reflex-search/reflex/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 6,019 This file is loaded in full into every session.
When invoked 6,019 The same file — it is already loaded in full.
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 $0.06019 $0.06019
Opus 5 $0.03010 $0.03010
Sonnet 5 $0.01204 $0.01204
Haiku 4.5 $0.00602 $0.00602

Measured 4d ago against content hash 4b394c66f4f5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

reflex 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 4d 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 · 614 lines

How it starts

The opening of the file, as written. The whole thing — 614 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

Project Overview

Reflex is a local-first, full-text code search engine written in Rust. It's a fast, deterministic replacement for Sourcegraph Code Search, designed specifically for AI coding workflows and automation.

Reflex uses trigram-based indexing to enable sub-100ms full-text search across large codebases (10k+ files). Unlike symbol-only tools, Reflex finds every occurrence of patterns—function calls, variable usage, comments, and more—not just definitions. Results include file paths, line numbers, and surrounding context, with optional symbol-aware filtering.


Core Principles

  1. Local-first: Runs fully offline; all data stays on the developer's machine
  2. Complete coverage: Finds every occurrence, not just symbol definitions
  3. Deterministic results: Same query → same answer; no probabilistic ranking
  4. Instant access: Trigram index + memory-mapping enables sub-100ms queries
  5. Agent-oriented: Clean JSON output built for AI coding agents and automation
  6. Regex support: Extract trigrams from patterns for fast regex search

Architecture Overview

Components

Module Description
Trigram Indexer Extracts trigrams from all code files; builds inverted index (trigram → file locations)
Content Store Stores full file contents (memory-mapped); enables context extraction around matches
Query Engine Intersects trigram posting lists; verifies matches; returns line-by-line results with context
Runtime Symbol Parser Uses Tree-sitter to parse candidate files at query time (only files matching trigrams)
Background Symbol Indexer Daemonized process that pre-caches symbols for faster queries on large codebases
Symbol Cache Persistent storage of parsed symbols (803-line caching system for instant symbol lookups)
CLI / API Layer Single binary for human and programmatic use (CLI and optional HTTP/MCP)
Watcher (optional) Incrementally updates index on file changes

Read the full file on GitHub · 614 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. 4d ago First seen · 614 lines · 6,019 tokens per session scan A 4b394c66f4f5

Subscribe to this mod's changes

reflex CLAUDE.md is an instructions file published in the GitHub repository reflex-search/reflex (74 stars, last pushed 1mo ago), licensed MIT. It adds 6,019 tokens to every session, about $0.0301 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.