harness-starter-kit AGENTS.md

harness-starter-kit AGENTS.md is an instructions file for Codex, OpenCode from harnessworks/harness-starter-kit. It costs 2,052 tokens per session, scanned A, original, MIT.

A set of project instructions for coding agents working with a reusable harness starter kit. It explains how agents should inspect and adapt an existing code repository while preserving its tools and conventions.

In plain words
What is it for?
Use it when adding durable rules, feedback loops, stored project knowledge, command routing, or checks that detect when a repository has drifted from its intended setup.
Why use it?
It prevents agents from blindly copying defaults, overwriting project files, or ignoring existing practices. It also encourages important rules to be checked through tests, linting, CI, or documented reviews.

Instructions file for CodexOpenCode

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/harnessworks/harness-starter-kit/agents-md
Clone the repo
git clone --depth 1 https://github.com/harnessworks/harness-starter-kit

Made for: Codex, OpenCode.

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 harness-starter-kit AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/harnessworks/harness-starter-kit/agents-md.svg)](https://agentmods.dev/instructions/harnessworks/harness-starter-kit/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/harnessworks/harness-starter-kit/agents-md"><img src="https://agentmods.dev/badge/instructions/harnessworks/harness-starter-kit/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,052 This file is loaded in full into every session.
When invoked 2,052 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 $0.02052 $0.02052
Opus 5 $0.01026 $0.01026
Sonnet 5 $0.00410 $0.00410
Haiku 4.5 $0.00205 $0.00205

Measured 4d ago against content hash 4231f92d2fdb, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

harness-starter-kit AGENTS.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.

AGENTS.md · 196 lines

How it starts

The opening of the file, as written. The whole thing — 196 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Purpose

This repository is a reusable harness engineering starter kit. It teaches agents how to add durable project rules, constraints, feedback loops, knowledge storage, and drift checks to another repository.

When applying this kit to a target repository, the target repository is the source of truth. Preserve its architecture, tools, package manager, docs, and conventions.

Core Rules

  • Keep this kit prompt-first. Agents should inspect the target repository and adapt the smallest useful harness pieces instead of copying defaults blindly.
  • Treat a target-local ./harness-starter-kit clone as read-only reference material unless the user explicitly asks to edit the kit itself.
  • Do not overwrite, delete, move, or re-clone target project files unless the user explicitly asks and the reason is clear.
  • Prefer the target repository's existing docs, scripts, CI, package manager, tests, and naming conventions over starter-kit defaults.
  • Make important rules enforceable where practical through lint, tests, type checks, import rules, CI, or drift checks. If automation is not practical, document the manual review point.
  • Keep templates generic and conservative. Do not bake in a single product architecture.
  • Record fixed user-visible runtime failures or high-risk bug paths that should not recur, including 5xx errors, crashes, security or permission bugs, data-loss risks, failed CI runs, failed harness checks, repeated agent mistakes, previously identified bug paths, or cross-environment mismatches in docs/failures/*.md unless the issue was purely transient or already covered by an existing failure note. If skipped, explain why in the final report.
  • When adding or updating a failure note, name the regression test, fixture, smoke check, lint rule, drift check, CI gate, or manual review point that prevents or detects recurrence. If no check is practical, record why.

Command Routing

  • /harness adopt: use commands/harness-adopt.md. Apply prompt-first harness engineering to the target repository. Inspect first, preserve target source-of-truth, avoid blind template copying, and finish with an adoption report.
  • /harness doctor: use commands/harness-doctor.md. It is diagnostic only: inspect and report, do not modify files, and do not remove a target-local ./harness-starter-kit directory.
  • /harness update: use commands/harness-update.md. Refresh the kit reference first, avoid blind overwrites, update .harness/source.json only after confirming the source, and finish with a Harness Update Report.
  • /harness refresh: use commands/harness-refresh.md. Review stale or duplicated target harness guidance. Do not delete, archive, move, or rename files without explicit approval for the specific files.
  • /harness review sub-agent: use commands/harness-review.md in sub-agent invocation mode. Treat the request as explicit permission to use a read-only reviewer subagent when available and permitted by the active runtime and tool instructions; if unavailable, blocked, not permitted, or failed, fall back to single-agent review and report the reason.
  • /harness review: use commands/harness-review.md. Review the current change set from an opposing harness-engineering perspective. It is diagnostic by default and must not modify files unless the user explicitly asks to apply fixes after seeing the review.

Read the full file on GitHub · 196 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 · 196 lines · 2,052 tokens per session scan A 4231f92d2fdb

Subscribe to this mod's changes

harness-starter-kit AGENTS.md is an instructions file published in the GitHub repository harnessworks/harness-starter-kit (113 stars, last pushed 2mo ago), licensed MIT. It adds 2,052 tokens to every session, about $0.0103 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.