sd0x-harness CLAUDE.md

sd0x-harness CLAUDE.md is an instructions file for coding agents from sd0xdev/sd0x-harness. It costs 2,000 tokens per session, scanned A, original, MIT.

A set of operating rules for Claude Code work, including required checks, automatic work loops, and limits on the changes an agent may make.

In plain words
What is it for?
Use it when developing with Claude Code to follow project rules, run required checks, handle stop-hook reminders, and manage automated development loops.
Why use it?
It gives the coding agent a defined way to decide what to check, when to continue, and when human input is allowed.

Instructions file

Installs and runs on its own, but its text points at files inside the plugin that ships it — anything it tells you to read at a ${CLAUDE_PLUGIN_ROOT} path is only there once the plugin is installed.

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/sd0xdev/sd0x-harness/claude-md
Clone the repo
git clone --depth 1 https://github.com/sd0xdev/sd0x-harness

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 sd0x-harness CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/sd0xdev/sd0x-harness/claude-md.svg)](https://agentmods.dev/instructions/sd0xdev/sd0x-harness/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/sd0xdev/sd0x-harness/claude-md"><img src="https://agentmods.dev/badge/instructions/sd0xdev/sd0x-harness/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,000 This file is loaded in full into every session.
When invoked 2,000 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 $0.02000 $0.02000
Opus 5 $0.01000 $0.01000
Sonnet 5 $0.00400 $0.00400
Haiku 4.5 $0.00200 $0.00200

Measured 4d ago against content hash f464621af2ea, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

sd0x-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 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 · 75 lines

How it starts

The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.

sd0x-dev-flow — Harness Engineering for Claude Code

How binding is a line in this file? Three tiers: Anchor (never deviate), Default (the normal call; deviate by stating a [DEVIATION] line that cites a fact signal, then keep working), Guidance (advisory). This file's own baseline is Default and lines above that baseline are marked inline — @rules/discretion.md classifies the plugin-managed rules/*.md, not this file, but its Anchor Register is the authority everywhere: a line here that hits the Register is Anchor no matter how it is worded or where it is restated.

Judgment inside the Default range is the expected behaviour, not a tolerated exception: decide from the change in front of you and continue. Uncertainty alone is not a reason to stop and ask — the human exits are the union of the ones enumerated in @rules/auto-loop.md and, for scope, in the contract @rules/scope-discipline.md points to (skills/codex-code-review/references/scope-contract.md § Human Exits); those two enumerations are the closed list, not this sentence.

Required Checks (Stop Hook reminded)

This table constrains the end state, not your choreography. How you batch edits, how deep you review, and when you run each gate are yours to choose; what is fixed is that every gate a change class requires has passed after the last edit in that class.

Change Type Must Run Can Skip
code files /codex-review-fast -> /precommit -
.md docs /codex-review-doc /codex-review-fast

Comment-only edits get no free pass: comments can carry compiler/lint/build directives (scripts/lib/utils.js:142 proves it), so edits to code files are conservatively classified as code even when only comments changed. Before PR: /pr-review.

The Stop Hook is a reminder, not a gate (hook-lightweighting, 2026-08-13): it prints which gates the reminder state still shows as owed and always exits 0. What binds is the behaviour layer — the terminal completion invariant in @rules/auto-loop.md — with verdicts recorded via node scripts/review-state.js note, digest-bound so an edit re-opens its plane. Details: docs/features/hook-lightweighting/2-tech-spec.md §3.

Read the full file on GitHub · 75 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 · 75 lines · 2,000 tokens per session scan A f464621af2ea

Subscribe to this mod's changes

sd0x-harness CLAUDE.md is an instructions file published in the GitHub repository sd0xdev/sd0x-harness (188 stars, last pushed yesterday), licensed MIT. It adds 2,000 tokens to every session, about $0.0100 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.