aios CLAUDE.md

Project instructions for AIOS that explain the difference between the AIOS installation folder and the project folder where commands are run.

In plain words
What is it for?
Resolving paths correctly, locating project state, and distinguishing framework scripts from files belonging to the current project.
Why use it?
They prevent the agent from reading or changing project files in the framework directory by mistake when both locations use relative paths.

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/rexleimo/aios/claude-md
Clone the repo
git clone --depth 1 https://github.com/rexleimo/aios
Per session 3,418 This file is loaded in full into every session.
When invoked 3,418 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 $0.03418 $0.03418
Opus 5 $0.01709 $0.01709
Sonnet 5 $0.00684 $0.00684
Haiku 4.5 $0.00342 $0.00342

Measured 3d ago against content hash fbb0a1ae1c45, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

aios 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.

Enumerates other installed skillsmediumAgent snooping

Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.

ls -la .aios/ .claude/skills/ docs/plans/
CLAUDE.md · 311 lines

How it starts

The opening of the file, as written. The whole thing — 311 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Path Conventions (路径约定)

IMPORTANT: All relative paths in this document are relative to the PROJECT ROOT (current working directory), NOT AIOS_ROOT.

Two Key Paths

  1. AIOS_ROOT / AIOS_ROOT_DIR - AIOS framework installation location

    • Contains scripts/, package.json, core framework code
    • Example: /Users/rex/.rexcil/aios or ~/aios-framework
    • Set via environment variable or auto-detected
  2. Project Root - Your current working directory where you run AIOS commands

    • Contains .aios/, .claude/skills/, docs/plans/, project-specific state
    • This is where pwd points when you run aios commands
    • All paths in this document like .aios/context-db/, .claude/skills/ are relative to this directory

Why This Matters

  • If your AIOS_ROOT=/Users/rex/.rexcil/aios but you're working in /Users/rex/my-project/:
    • .aios/context-db/ means /Users/rex/my-project/.aios/context-db/ (✓ correct)
    • NOT /Users/rex/.rexcil/aios/.aios/context-db/ (✗ wrong)
  • Framework scripts live in $AIOS_ROOT/scripts/
  • Your project state lives in $(pwd)/.aios/

Verification Commands

# Check your project root
pwd

# Check AIOS framework location
echo $AIOS_ROOT

# List your project's AIOS directories
ls -la .aios/ .claude/skills/ docs/plans/

Evaluate the AIOS workflow policy before creating a plan, selecting a skill, or dispatching work.

  1. direct and noop turns do not create a persistent plan or invoke a skill chain.
  2. guarded turns use only the required edit and verification gates.
  3. planned turns create or reuse one work-item plan, then execute only the Provider selected by the current rex Command.
  4. Do not insert a global bootstrap or compatibility skill chain. aios-workflow-router helps classify work but does not replace the Provider selected by the current Rex Capability Command.
  5. Invoke the selected Provider rather than paraphrasing it; use verification-before-completion before claiming a changed behavior is complete.

Read the full file on GitHub · 311 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 · 311 lines · 3,418 tokens per session scan B fbb0a1ae1c45

Subscribe to this mod's changes

aios CLAUDE.md is an instructions file published in the GitHub repository rexleimo/aios (52 stars, last pushed 4d ago), licensed MIT. It adds 3,418 tokens to every session, about $0.0171 per session on Opus 5. A static security scan graded it B with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.