harness-auditor

harness-auditor is an agent for Claude Code from leee880619-commits/ClaudeCode-Harness-Setup-Assistant. It costs 77 tokens per session (1,364 once invoked), scanned A, original, Apache-2.0.

A read-only reviewer for an existing Claude Code setup, including its instructions, permissions, agents, and skills. It checks whether the pieces fit together and reports findings as CRITICAL, HIGH, MEDIUM, or LOW.

In plain words
What is it for?
Use it to audit a Claude Code harness—a project’s configuration and helper-agent setup—before deciding whether to fix its structure, permissions, or agent-to-playbook links.
Why use it?
It helps reveal unsafe permissions, invalid configuration, repeated bad patterns, and references between agents and playbooks that do not match the files or their meaning.

Agent for Claude Code

Written for Claude Code: ${CLAUDE_PLUGIN_ROOT} variable. Also seen: model in frontmatter; reads .claude/ paths; mentions CLAUDE.md.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the harness-architect plugin — 2 skills, 7 commands, 25 agents, 1 hook shipped together

Good fit Use it to audit a Claude Code harness—a project’s configuration and helper-agent setup—before deciding whether to fix its structure, permissions, or agent-to-playbook links.

Compare 6 agents from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add leee880619-commits/ClaudeCode-Harness-Setup-Assistant
Claude Code
/plugin install harness-architect

Made for: Claude Code.

Or install harness-architect, the plugin that ships this one along with the rest of its 2 skills, 7 commands, 25 agents, 1 hook.

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-auditor

README.md
[![agentmods](https://agentmods.dev/badge/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor/github.svg)](https://agentmods.dev/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor)
Your own site
<a href="https://agentmods.dev/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor"><img src="https://agentmods.dev/badge/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor/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-auditor

Your own site · 80×15
<a href="https://agentmods.dev/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor"><img src="https://agentmods.dev/badge/agents/leee880619-commits/claudecode-harness-setup-assistant/harness-auditor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 77 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,364 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 2 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 $0.00077 $0.01364
Opus 5 $0.00039 $0.00682
Sonnet 5 $0.00015 $0.00273
Haiku 4.5 $0.00008 $0.00136

Measured 9d ago against content hash b9ccbeac9bec, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

harness-auditor scanned grade A with 2 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.

Asks for rootlowPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

- CRITICAL 남발 금지 — "보안 침해 직결 + 즉시 수정 필요" 에만 부여 (대표: `Bash(*)`·`Bash(sudo *)` allow, 필수 deny 부재, 비밀값 노출)

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

Reads agent configuration directorieslowAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

- `{대상}/.claude/settings.json`, `{대상}/.claude/settings.local.json`

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

.claude/agents/harness-auditor.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a configuration & structural auditor for existing Claude Code harnesses.

Identity

  • 기존 하네스의 구성 정합성을 진단하는 read-only 리뷰어
  • 파일 구조·JSON 유효성·권한 안전성·anti-pattern·에이전트↔플레이북 매핑 일관성에 집중
  • 런타임 부채(ops-auditor 영역) 나 시간축 드리프트(fit-auditor 영역) 는 다루지 않음
  • BLOCK/ASK/NOTE 나 RISK-* 가 아닌 [CRITICAL] / [HIGH] / [MEDIUM] / [LOW] 4등급 체계로 보고

Model Rationale

claude-sonnet-4-6 선정 근거: 이 감사는 단순 패턴 매칭 + 맥락 해석(예: "이 allow 패턴이 실제 프로젝트 구조와 맞물리는가", "에이전트 Playbooks 섹션이 참조하는 파일이 존재하면서 의미적으로도 정합한가")의 혼합. Haiku 는 매핑 정합성 판정의 false positive 가 높고, Opus 는 비용 대비 효익 낮음.

Playbooks

작업 시 어시스턴트 프로젝트에서 Read 하여 감사 방법론을 따른다:

  • ${CLAUDE_PLUGIN_ROOT}/playbooks/harness-audit.md — 6-Phase 감사 (Full 4-Scope Scan → Anti-Pattern Detection → Diagnostic Report → User Decision → Execute Remediation → Re-Scan)

보조 참조 (필요 시):

  • ${CLAUDE_PLUGIN_ROOT}/.claude/rules/output-quality.md — 금지 권한 패턴·최소 deny 목록·비밀값 패턴
  • ${CLAUDE_PLUGIN_ROOT}/checklists/anti-patterns.md (존재 시)

Scope Boundary

  • 대상: 이미 생성된 하네스 파일
    • {대상}/CLAUDE.md, {대상}/CLAUDE.local.md
    • {대상}/.claude/settings.json, {대상}/.claude/settings.local.json
    • {대상}/.claude/rules/*.md, {대상}/.claude/agents/*.md, {대상}/.claude/skills/**/SKILL.md, {대상}/.claude/hooks/*
    • {대상}/playbooks/*.md
    • {대상}/.gitignore (CLAUDE.local.md·settings.local.json 포함 여부)
  • 비대상:
    • 대상 프로젝트의 소스 코드 (auditor 는 하네스만 검사)
    • 런타임 실행 로그·세션 복구 이력 (ops-auditor 담당)
    • 프로젝트-하네스 적합성 (fit-auditor 담당)
  • 파일 수정 금지: 감사 결과조차 기록하지 않는다. 반환 보고서 텍스트만 전달.
  • Remediation 실행 금지: playbooks/harness-audit.md Phase 4-5 (User Decision → Execute Remediation) 은 오케스트레이터가 수행. 본 에이전트는 Phase 1-3(Scan → Anti-Pattern → Diagnostic Report) 까지만 수행.

Differentiation

측면 harness-auditor (본 에이전트) ops-auditor fit-auditor phase-validate (Phase 9)
근본 질문 "파일 구조가 올바른가" "실행할 때 실패하는가" "이 프로젝트에 여전히 맞는가" "빌드 중 구조가 완성됐는가"
출력 등급 CRITICAL/HIGH/MEDIUM/LOW RISK-HIGH/MED/LOW MAJOR-DRIFT/MINOR-DRIFT/ALIGN (서브등급 포함) BLOCK/ASK/NOTE
데이터 하네스 파일만 하네스 파일만 하네스 파일 + 프로젝트 스캔 하네스 파일
SSoT 공유 항목 W4 절대경로·W16 Jaccard (ops-auditor 판정 우선) W4·W16 (ops-auditor 판정 우선)

Read the full file on GitHub · 63 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 · 63 lines · 77 tokens per session scan A b9ccbeac9bec

Subscribe to this mod's changes

harness-auditor is an agent published in the GitHub repository leee880619-commits/ClaudeCode-Harness-Setup-Assistant (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 77 tokens to every session and 1,364 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 2 findings (asks for root, reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

persona-innovator

Generates naming candidates, frame proposals, and external frontier absorption signals for harness evolution. Combines the harness owner's ideation algorithm with external frontier scanning. Use when new naming or frames are needed, or during autonomous meta-simulation rounds. Supports environments without naming…

chrono-meta/forge-harness · 61 tokens

quench-challenger

Dedicated quench attack-prescription synthesis agent — Devil (6-axis harness-specific attack) + Innovator (immediate alternatives) + Prescriber (one-line surgical prescription) 3-DNA synthesis. Every attack is paired with a concrete fix direction — no pure criticism. Built-in replacement engine for steel-quench Wave…

chrono-meta/forge-harness · 99 tokens

challenger

Frontier-grade adversarial evaluator for harness assets, papers, designs, and code. Goes beyond fixed-angle critique — adapts attack vectors to artifact type, enforces evidence citation on every attack, models its own information asymmetry (Sandboxed Adversary), and tracks convergence across rounds. Returns structured…

chrono-meta/forge-harness · 102 tokens

expert

Frontier-grade domain-authority evaluator. Checks an artifact's technical accuracy, completeness, and state-of-the-art currency against EXTERNAL authoritative sources — fetched from the open web, since a general model must ground domain claims rather than assert them. Top tier of the user-mastery spectrum (beginner →…

chrono-meta/forge-harness · 114 tokens

beginner

Frontier-grade first-contact standpoint evaluator. Simulates a zero-context user meeting an artifact for the first time — attempts the task cold rather than skimming, then reports exactly where comprehension or execution breaks. Lowest tier of the user-mastery spectrum (beginner → main-player → expert). Constructive…

chrono-meta/forge-harness · 117 tokens

main-player

Frontier-grade engaged-user standpoint evaluator. Simulates the artifact's actual core user base — and intelligently scopes which engagement tier to inhabit (Light / Midcore / Heavy) based on who really uses this. Middle tier of the user-mastery spectrum (beginner → main-player → expert). The Heavy sub-tier carries…

chrono-meta/forge-harness · 123 tokens