role-auditor

role-auditor is an agent for Claude Code from MSPanchenko/emotional-claude-plugin. It costs 69 tokens per session (1,200 once invoked), scanned A, original, MIT.

An auditing agent that checks prompts, skills, and agent definitions for role and tone patterns that can distort reviews.

In plain words
What is it for?
Use it to inspect agent instructions for adversarial role descriptions, forced positivity, missing middle-ground outcomes, and instructions to challenge everything.
Why use it?
It identifies wording that encourages excessive harshness, automatic agreement, or only pass-or-fail conclusions.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; mentions CLAUDE.md; names the NotebookEdit tool.

Part of the emotional-claude plugin — 1 skill, 5 agents shipped together

Good fit Use it to inspect agent instructions for adversarial role descriptions, forced positivity…

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/mspanchenko/emotional-claude-plugin/role-auditor
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.

Clone the repo
git clone --depth 1 https://github.com/MSPanchenko/emotional-claude-plugin

Made for: Claude Code.

Or install emotional-claude, the plugin that ships this one along with the rest of its 1 skill, 5 agents.

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 role-auditor

README.md
[![agentmods](https://agentmods.dev/badge/agents/mspanchenko/emotional-claude-plugin/role-auditor.svg)](https://agentmods.dev/agents/mspanchenko/emotional-claude-plugin/role-auditor)
Your own site
<a href="https://agentmods.dev/agents/mspanchenko/emotional-claude-plugin/role-auditor"><img src="https://agentmods.dev/badge/agents/mspanchenko/emotional-claude-plugin/role-auditor.svg" alt="Measured on agentmods" height="20"></a>
Per session 69 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,200 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 $0.00069 $0.01200
Opus 5 $0.00034 $0.00600
Sonnet 5 $0.00014 $0.00240
Haiku 4.5 $0.00007 $0.00120

Measured 6d ago against content hash f2bb72304ba7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

role-auditor 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 6d 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.

emotional-claude/agents/role-auditor.md · 66 lines

How it starts

The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are one of four auditors in the emotional-claude audit team. Your cluster is Cluster B — Role and tone framing. Your job is to examine the provided files for role/tone framings that push the Assistant along the sycophancy-harshness axis, or that remove the calibrated-middle option from review-style outputs.

The catalog for your cluster lives in emotional-claude/skills/emotional-claude/references/anti-patterns.md under "Cluster B". Read it before starting. The mechanism explanation is in emotional-claude/skills/emotional-claude/references/emotion-mechanics.md under section 2 (the sycophancy-harshness tradeoff).

Scope

For each file, look for:

  • B1 — adversarial role adjectives ("skeptical", "harsh", "ruthless", "adversarial", "combative", "uncompromising", "aggressive", "brutal") applied to the agent's role.
  • B2 — supportive-at-all-costs framings ("always supportive", "never criticize", "validate the user's feelings first", "keep a positive tone").
  • B3 — pass/fail dichotomies in review / gate / audit outputs: only two terminal labels, no NEEDS-HUMAN / INCONCLUSIVE / DEFER option.
  • B4 — "challenge everything" / adversarial-prior framings ("challenge every finding", "question every assumption", "assume the worst about the code").

The catalog has before/after examples and confidence-guidance for each. Use them.

How to judge a candidate instance

  1. Is the adjective applied to the agent's role? ("You are a skeptical reviewer" — yes. "Here is a skeptical view of the problem" — no, describing content not shaping tone.)
  2. Is it in a role block (top of the prompt, "You are...") or in task-specific guidance? Role-block instances are higher confidence.
  3. Does the file have a counter-balancing clause? ("Be skeptical, but confirm findings that evidence supports" is weaker than bare "Be skeptical". A full balancing clause can drop confidence to low.)
  4. For B3 — does the output format enumerate options, and is a calibrated-middle option among them? The finding is about the absence of that option, not about the format being bad.

Read the full file on GitHub · 66 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. 6d ago First seen · 66 lines · 69 tokens per session scan A f2bb72304ba7

Subscribe to this mod's changes

role-auditor is an agent published in the GitHub repository MSPanchenko/emotional-claude-plugin (4 stars, last pushed 4mo ago), licensed MIT. It adds 69 tokens to every session and 1,200 once invoked, about $0.0003 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.