groundtruth: Instructions file for Claude Code

CLAUDE.md

groundtruth CLAUDE.md is an instructions file for Claude Code from akahkhanna/groundtruth. It costs 2,531 tokens per session, scanned A, original, MIT.

Repository instructions for Groundtruth, a Claude Code plugin that checks whether an agent's claimed work matches the actual code changes and project rules.

In plain words
What is it for?
Use it to verify claims against git changes and transcripts, and to check for issues such as secrets, placeholder code, dropped symbols, and rule violations.
Why use it?
It helps catch unfinished, undeclared, or incorrectly reported work before a coding task is marked done, using local checks rather than network services or language-model calls.

Instructions file for Claude Code

Written for Claude Code: SessionStart hook event. Also seen: mentions CLAUDE.md; mentions Claude Code.

This is akahkhanna/groundtruth's own configuration. It tells Claude Code how to work on groundtruth itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything groundtruth configures →

Reuse

Borrowing it

Nothing to install: this file belongs to akahkhanna/groundtruth. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/akahkhanna/groundtruth/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/akahkhanna/groundtruth

Made for: Claude Code.

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 groundtruth CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/akahkhanna/groundtruth/claude-md.svg)](https://agentmods.dev/instructions/akahkhanna/groundtruth/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/akahkhanna/groundtruth/claude-md"><img src="https://agentmods.dev/badge/instructions/akahkhanna/groundtruth/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,531 This file is loaded in full into every session.
When invoked 2,531 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.02531 $0.02531
Opus 5 $0.01265 $0.01265
Sonnet 5 $0.00506 $0.00506
Haiku 4.5 $0.00253 $0.00253

Measured 7d ago against content hash 180a0fad2669, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

groundtruth 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 7d 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 · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 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.

What this is

Groundtruth is a Claude Code plugin (distributed via the plugin marketplace, not npm) that catches the false "Done." It's a deterministic local hook — no LLM calls, no network, no API key. It reads an agent turn from the outside (the request, the claim, the real git diff, the project rules) and renders a one-screen verdict before the turn ends.

As of v2.0.0 the default honesty/completeness engine is the "claims contract" (declare-then-verify) in hooks/claims-contract.mjs: the agent ends a code-changing turn with one fenced groundtruth-claims block, and the engine verifies every claim against the diff/transcript (CA — claimed-but-absent), every agent-authored change against the claims (UC — undeclared), and a missing/invalid block on authored work (NC). This replaced the old prose-parsing layer (v1 class-1/class-3, retired). Opt back out with GROUNDTRUTH_CONTRACT=0. The diff-facing checks (secrets, RLS, stubs, dropped symbols, compiled rules, Rule Zero) are unchanged.

Design invariant, enforced in review: a false positive is treated as fatal. A check earns its place only if it's provably correct in a defined scope. Outside that scope a check must abstain or degrade to a bounded warn — never a clean green. When adding or changing a check, this is the bar. See CONTRIBUTING.md.

Commands

node hooks/groundtruth.test.mjs      # 479 engine unit checks, no deps, no framework
node hooks/claims-contract.test.mjs  # 217 v2 claims-contract checks (same assert style)
node hooks/redteam.mjs               # live adversarial harness (37 checks across sandboxed throwaway repos)
npm test                             # runs BOTH unit files (479 engine + 217 contract = 696)
npm run check-self                   # engine checks, but echoes the exit code as a GREEN/RED done-verdict

The test files are a flat sequence of ok(label, cond) asserts — there is no single-test runner; you run the whole file (it's fast). They import the pure exports of groundtruth.mjs/claims-contract.mjs/compile-rules.mjs/symbol-integrity.mjs directly, so a new check is only testable if it's exported and pure. Add a regression test in the same style next to the code it covers, and log the fix in FIXES.md (symptom → root cause → fix → regression test).

Read the full file on GitHub · 74 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. 7d ago First seen · 74 lines · 2,531 tokens per session scan A 180a0fad2669

Subscribe to this mod's changes

groundtruth CLAUDE.md is an instructions file published in the GitHub repository akahkhanna/groundtruth (6 stars, last pushed 1mo ago), licensed MIT. It adds 2,531 tokens to every session, about $0.0127 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.

Related

Other instructions, from other repositories

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

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).

microsoft/vscode · 5,001 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,469 tokens

spec-kit AGENTS.md

AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,104 tokens