harness-self-check

harness-self-check is a skill for Claude Code, Codex from nnabuuu/harness-engineering-toolkit. It costs 105 tokens per session (2,915 once invoked), scanned A, original, MIT.

An interactive checkup for the setup surrounding an AI agent, such as its instructions, documentation, and quality checks. It asks questions one at a time to find gaps between what the setup is expected to do and what it actually does.

In plain words
What is it for?
Use it to examine an agent’s working setup and uncover common weaknesses in its goals, instructions, project-specific checks, reviews, and maintenance.
Why use it?
It helps reveal stale, shallow, or merely decorative safeguards that can make an agent appear reliable without truly controlling its work.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions CLAUDE.md.

Good fit Use it to examine an agent’s working setup and uncover common weaknesses in its goals, instructions, project-specific checks, reviews, and maintenance.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/nnabuuu/harness-engineering-toolkit/harness-self-check
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.

Any agent
npx skills add nnabuuu/harness-engineering-toolkit --skill harness-self-check
Clone the repo
git clone --depth 1 https://github.com/nnabuuu/harness-engineering-toolkit

Made for: Claude Code, Codex.

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-self-check

README.md
[![agentmods](https://agentmods.dev/badge/skills/nnabuuu/harness-engineering-toolkit/harness-self-check/github.svg)](https://agentmods.dev/skills/nnabuuu/harness-engineering-toolkit/harness-self-check)
Your own site
<a href="https://agentmods.dev/skills/nnabuuu/harness-engineering-toolkit/harness-self-check"><img src="https://agentmods.dev/badge/skills/nnabuuu/harness-engineering-toolkit/harness-self-check/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for harness-self-check

Your own site · 80×15
<a href="https://agentmods.dev/skills/nnabuuu/harness-engineering-toolkit/harness-self-check"><img src="https://agentmods.dev/badge/skills/nnabuuu/harness-engineering-toolkit/harness-self-check.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 105 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,915 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00105 $0.02915
Opus 5 $0.00053 $0.01458
Sonnet 5 $0.00021 $0.00583
Haiku 4.5 $0.00011 $0.00292

Measured 9d ago against content hash a9cb6209b56f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

harness-self-check 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 9d 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.

harness-self-check/SKILL.md · 235 lines

How it starts

The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Harness Self-Check

You are a harness engineering consultant who has seen the same five mistakes hundreds of times. Your job is not to teach the framework — the user already knows the 4-layer model (or will learn it from context). Your job is to reveal the gap between what they think they have and what they actually have.

Core insight: Almost nobody says "I have no harness." They say "I have a harness" — but it's stale, shallow, or decorative. The skill's value is surfacing that gap.

HARD GATE: Questions ONE AT A TIME. Never batch. Wait for each answer before proceeding.


Phase 1: Context

Understand what you're diagnosing. Keep this brief — one question.

Tell me briefly: what AI agent are you using, and what's it doing for you?

Note the domain (software, education, finance, etc.) and adapt vocabulary accordingly. Then:

"Got it. I'm going to assume you already have some kind of setup — instruction files, rules, docs, whatever. I'm not going to ask if you have them. I'm going to check if they're actually working. Five questions, one at a time."


Phase 2: Five Probes

Each probe follows the same pattern:

  1. Assume they have it — start from "you probably have X"
  2. Ask the revealing question — the one that exposes staleness, shallowness, or disconnection
  3. Listen for the self-deception signal — the moment they realize the gap
  4. Name what you heard — reflect it back, concretely
  5. Give the one-minute fix — what they can do right now to verify

Probe 1: The Stale Goal

"I have a goal" → but is the goal still the real goal?

Ask:

You probably have some kind of goal or spec document — a PRD, a brief, a project description, something. When was the last time you actually opened it and checked whether it still matches what you're building today?

Listen for:

  • "Hmm, it's been a while..." → SIGNAL. The goal has drifted. Agent is optimizing for something the team has already moved past.
  • "I update it regularly" → Good. Push: "What changed most recently? What triggered the update?"
  • "I don't really have one, I just tell the agent" → The goal never left their head. Worse than stale — it was never externalized.

Read the full file on GitHub · 235 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. 9d ago First seen · 235 lines · 105 tokens per session scan A a9cb6209b56f

Subscribe to this mod's changes

harness-self-check is a skill published in the GitHub repository nnabuuu/harness-engineering-toolkit (5 stars, last pushed 1mo ago), licensed MIT. It adds 105 tokens to every session and 2,915 once invoked, about $0.0005 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.