What is harness CLAUDE.md?

Project guidance for a standalone runtime that runs AI agents through a terminal interface or a web interface. It provides shared agent loading, tools, sub-agents, sessions, memory, and several access interfaces.

Orienting yourself in the codebase, running the terminal or web application, working with agents and tools, and finding the relevant TypeScript, React, Fastify, and configuration files.

It helps contributors understand where the runtime, tools, web application, terminal interface, configuration, and authentication code belong. This reduces guesswork when changing the project.

harness CLAUDE.md — Instructions file by mastersof-ai

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/mastersof-ai/harness/claude-md

Clone the repo

git clone --depth 1 https://github.com/mastersof-ai/harness

Per session 798 This file is loaded in full into every session.

When invoked 798 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.

Model	Per session	Once invoked
Fable 5	$0.00798	$0.00798
Opus 5	$0.00399	$0.00399
Sonnet 5	$0.00160	$0.00160
Haiku 4.5	$0.00080	$0.00080

Measured 2d ago against content hash 40995781a7c1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

harness 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 2d 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 · 57 lines

How it starts

The opening of the file, as written. The whole thing — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Masters Of AI Harness

DESIGN.md — overview with links to docs/
docs/ — architecture, agents, tools, configuration, sandbox, design decisions
CHANGELOG.md — version history

Quick Orientation

Standalone agent runtime built on top of the Claude Agent SDK. Two interfaces, one runtime:

Terminal TUI — mastersof-ai [--agent x] — React/Ink, single user, local
Web UI — mastersof-ai --serve — Fastify backend + React SPA frontend, multi-user, token auth

Both share agent loading, tools, sub-agents, sessions, and memory.

TypeScript, runs via tsx (no build step for backend)
Entry: bin/mastersof-ai.js → src/index.tsx
Agent loading: src/agent-context.ts + src/manifest.ts (frontmatter) → src/agent.ts
Tools: src/tools/ — in-process MCP servers, one per domain
Serve mode: src/serve.ts — Fastify HTTP/WS server (REST API + WebSocket streaming)
Web frontend: web/ — React + Vite + Tailwind SPA (deploys to Cloudflare Pages)
A2A: src/a2a/ — agent card generation, A2A protocol server/client
TUI: src/components/ — React/Ink (DO NOT TOUCH unless broken)
Config: ~/.mastersof-ai/config.yaml
Auth (serve mode): ~/.mastersof-ai/access.yaml
Security: src/env-safety.ts, src/url-safety.ts, src/content-safety.ts
Credentials: src/credentials.ts + src/egress-proxy.ts
Process isolation: src/ipc-protocol.ts, src/session-worker.ts, src/worker-manager.ts
WS protocol: src/ws-protocol.ts, src/sdk-stream.ts, src/query-mutex.ts
Health/observability: src/health.ts, src/rate-limit.ts, src/cost.ts
Privacy: src/privacy.ts — LGPD compliance
CLI subcommands: src/cli/ — modular command handlers

Running Locally

npx tsx bin/mastersof-ai.js                         # TUI with default agent
npx tsx bin/mastersof-ai.js --agent researcher      # TUI with specific agent
npx tsx bin/mastersof-ai.js --serve                 # Web UI server on port 3200
npx tsx bin/mastersof-ai.js --serve --port 5000     # Web UI on custom port
npx tsx bin/mastersof-ai.js --agent researcher --sandbox  # Bubblewrap sandbox
npx tsx bin/mastersof-ai.js --card                  # Output Agent Card JSON
npx tsx bin/mastersof-ai.js --list-agents

# CLI subcommands
npx tsx bin/mastersof-ai.js run billing "Run monthly billing"           # Headless run
npx tsx bin/mastersof-ai.js credentials check --agent billing           # Check agent credentials
npx tsx bin/mastersof-ai.js credentials migrate billing                 # Generate migration YAML
npx tsx bin/mastersof-ai.js access create --name partner --agents a,b   # Create access token
npx tsx bin/mastersof-ai.js access rotate --name partner                # Rotate token
npx tsx bin/mastersof-ai.js status billing                              # Recent run results
npx tsx bin/mastersof-ai.js preflight --agent billing                   # Validate full config
npx tsx bin/mastersof-ai.js create my-agent                             # Scaffold new agent

Read the full file on GitHub · 57 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.

2d ago First seen · 57 lines · 798 tokens per session scan A 40995781a7c1

Subscribe to this mod's changes

harness CLAUDE.md is an instructions file published in the GitHub repository mastersof-ai/harness (40 stars, last pushed 5mo ago), licensed MIT. It adds 798 tokens to every session, about $0.0040 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.

Data and AI Frontend LLM Memory and context a2a-protocol agent-runtime agent-sdk ai-agent claude

agentic-sandbox CLAUDE.md

Claude Code instructions for jmagly/agentic-sandbox, covering claude.md, repository purpose, tech stack, development commands and management server.

jmagly/agentic-sandbox · 2,801 tokens

agentic-sandbox AGENTS.md

AGENTS.md instructions for jmagly/agentic-sandbox, covering agents.md, project overview, tech stack, development commands and project structure.

jmagly/agentic-sandbox · 2,449 tokens

Sponsio CLAUDE.md

Instructions for SponsioLabs/Sponsio, covering agent guide for sponsio, what sponsio is, positioning, what to read first and repository map.

SponsioLabs/Sponsio · 1,937 tokens

gini-agent AGENTS.md

AGENTS.md instructions for Open-Curiosity/gini-agent, covering gini agent instructions, shape, adrs, boundaries and branches.

Open-Curiosity/gini-agent · 3,639 tokens

innovation-lab-examples typescript.instructions.md

Instructions for fetchai/innovation-lab-examples, a project described as: 80+ production-ready AI agent examples in Python — build autonomous agents, multi-agent systems and agentic AI with uAgents, ASI:One, MCP, A2A, LangChain, CrewAI, Gemini, Claude and OpenAI.

fetchai/innovation-lab-examples · 12 tokens

innovation-lab-examples python.instructions.md