What is agentware CLAUDE.md?

Project instructions for Agentware, a workspace that loads general guidance from a separate knowledge directory. They define a startup check that must confirm the workspace has been initialized before other work begins.

Use them when starting tasks in Agentware, checking its configuration, completing onboarding, or working with its session startup rules.

They prevent work from starting without the required external instructions and configuration. The setup keeps personal knowledge outside the repository.

agentware CLAUDE.md — Instructions file for coding agents by r5rana: Project instructions for Agentware, a workspace that loads general…

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/r5rana/agentware/claude-md

Clone the repo

git clone --depth 1 https://github.com/r5rana/agentware

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/r5rana/agentware/claude-md.svg)](https://agentmods.dev/instructions/r5rana/agentware/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/r5rana/agentware/claude-md"><img src="https://agentmods.dev/badge/instructions/r5rana/agentware/claude-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 624 This file is loaded in full into every session.

When invoked 624 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.

Model	Per session	Once invoked
Fable 5	$0.00624	$0.00624
Opus 5	$0.00312	$0.00312
Sonnet 5	$0.00125	$0.00125
Haiku 4.5	$0.00062	$0.00062

Measured 5d ago against content hash c12446e88861, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

agentware 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 5d 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.

- IF the knowledge dir is unconfigured OR its .initialized sentinel is absent THEN RUN the onboarding skill in .claude/skills/onboarding/SKILL.md before any other work. [R-BOOT-01]

CLAUDE.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

agentware — Claude Code project instructions

This is an agentware workspace: a clone-and-go steering framework. This repo holds ONLY generic steering. The operator's knowledge base lives in an EXTERNAL directory chosen at onboarding — nothing personal is ever committed here.

Claude Code auto-loads this file on every session. It pulls in the canonical methodology and steering below via imports, so those rules are always in context.

🔴 Bootstrap gate — check FIRST, before any task

The external knowledge directory is resolved with scripts/agentware config --knowledge-dir-only. The SessionStart hook prints an AGENTWARE_STATUS: line each session reporting whether it is initialized.

IF the knowledge dir is unconfigured OR its .initialized sentinel is absent THEN RUN the onboarding skill in .claude/skills/onboarding/SKILL.md before any other work. [R-BOOT-01]
IF onboarding completes THEN write the .initialized sentinel (timestamp + handle) in the knowledge dir and resume the original request. [R-BOOT-02]
IF the knowledge dir is configured AND initialized THEN proceed with the request normally. [R-BOOT-03]
NEVER create the .initialized sentinel without actually running onboarding. [R-BOOT-04]
NEVER skip onboarding because the user gave a quick task; run it once, then do the task. [R-BOOT-05]

Canonical methodology (imported, always-on)

@AGENTS.md

Always-on steering (imported)

@steering/common-problems.md @steering/project-context.md

Working in this repo

The 3-phase loop is ./agentware.sh <feature> (see docs/loop.md).
The deterministic toolkit is scripts/agentware — the ONLY writer of the knowledge index. Resolve the external knowledge dir at runtime; never hardcode it.
The roles are two Claude Code subagents in .claude/agents/: agentware-planner (plan only, never executes) and agentware-execution (implements the loop; the loop's POST phase self-assesses via this agent).
Skills live in .claude/skills/ — the loop skills (onboarding, knowledge-base, self-improvement, ui-verification) plus the 14 authored default skills (6 security, 6 productivity, 2 gap-fillers) catalogued in .claude/skills/README.md and catalog/skills.json; manage them with scripts/agentware skill list/search/add/remove/validate. The /agentware-plan command scaffolds a feature plan.
On failure, follow the failure-handling escalation ladder in AGENTS.md (R-FAIL-01..08): KB → own reasoning → change inputs → switch approach (after ≤3 tries) → web search. Keep moving; never re-loop a tier.

Read the full file on GitHub · 48 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.

5d ago First seen · 48 lines · 624 tokens per session scan B c12446e88861

Subscribe to this mod's changes

agentware CLAUDE.md is an instructions file published in the GitHub repository r5rana/agentware (24 stars, last pushed 18d ago), licensed Apache-2.0. It adds 624 tokens to every session, about $0.0031 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.

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

agentware CLAUDE.md