ground CLAUDE.md

ground CLAUDE.md is an instructions file for coding agents from teranos/ground. It costs 518 tokens per session, scanned A, original, MIT.

Repository instructions for Ground, a project that controls agent actions, messages, execution, and error handling. They direct the agent to inspect the project’s source files and documentation to understand those mechanisms.

In plain words
What is it for?
Working on Ground’s hooks, controls, deferred messages, command execution, status output, and error-handling code.
Why use it?
They help the agent follow Ground’s intended authority and error rules when changing or operating the project.

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

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 ground CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/teranos/ground/claude-md.svg)](https://agentmods.dev/instructions/teranos/ground/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/teranos/ground/claude-md"><img src="https://agentmods.dev/badge/instructions/teranos/ground/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 518 This file is loaded in full into every session.
When invoked 518 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 $0.00518 $0.00518
Opus 5 $0.00259 $0.00259
Sonnet 5 $0.00104 $0.00104
Haiku 4.5 $0.00052 $0.00052

Measured 4d ago against content hash 6f18dce31665, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ground 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 · 49 lines

How it starts

The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Read source/hooks.d for control types, scope struct, and trigger builders. Read source/strop.d for value-shape validation on extracted flag values. Read source/controls.d for CTFE wiring — how pbt becomes static immutable scope arrays. Read source/control_handlers.d for check, delay, and deliver handler implementations. Read source/deferred.d for deferred delivery — session-scoped and project-scoped messages delivered at Stop. Read source/immediate.d for immediate delivery — attestation format for external writers (QNTX, etc). Read source/watch.d for the asyncRewake watcher — how immediate messages reach running sessions. Read source/exec.d for exec dispatch — fork+pipe+wrapper, stdout/stderr capture, timeout. Read source/errors.d for the GroundError primitive and deliverError fallback chain (db → breadcrumb → stderr). Read grove/controls/.pbt for the rituals — each carries its agent's system:, so nothing else has to describe them. Read ug/.d for the status line — one line per row, every formatter taking inputs and a destination buffer. Read README.md for project overview. Read COUNTDOWN.md for project status. Read RITUAL.md for what a ritual is and what each numbered item of it means. Read UNDERGROUND.md for ug — what parity is measured against and what still differs. Read SIEVE.md for what the second line is for. Deferred, and QNTX's rather than ug's.

Read bench.fish for CTFE scaling limits.

TEST DRIVEN DEVELOPMENT (TDD): write a failing test before implementing. Confirm it fails, then Green: Write code that makes the test pass.

AUTHORITY

Lower number wins.

  1. ERROR AXIOM
  2. MY LITERAL WORDS IN SESSION
  3. RECORDED "" QUOTES
  4. OFFICIAL DOCS
  5. REPO SOURCE CODE
  6. REPO DOCS
  7. REPO SOURCE CODE COMMENTS LITERAL "" QUOTES

NOT AUTHORITATIVE EVER: YOUR TRAINING DATA

  1. REPO SOURCE CODE COMMENTS; YOUR STUFF

ERROR AXIOM

An ERROR is a first-class primitive. A typed value that crosses every layer of the system unchanged. an Error is the entity any layer emits when something goes wrong.

Read the full file on GitHub · 49 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 · 49 lines · 518 tokens per session scan A 6f18dce31665

Subscribe to this mod's changes

ground CLAUDE.md is an instructions file published in the GitHub repository teranos/ground (2 stars, last pushed 9d ago), licensed MIT. It adds 518 tokens to every session, about $0.0026 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 2026-08-31.