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.
git clone --depth 1 https://github.com/queelius/claude-anvilWrote 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.
[](https://agentmods.dev/agents/queelius/claude-anvil/voice-auditor)<a href="https://agentmods.dev/agents/queelius/claude-anvil/voice-auditor"><img src="https://agentmods.dev/badge/agents/queelius/claude-anvil/voice-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.
<a href="https://agentmods.dev/agents/queelius/claude-anvil/voice-auditor"><img src="https://agentmods.dev/badge/agents/queelius/claude-anvil/voice-auditor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00043 | $0.02758 |
| Opus 5.5 | $0.00017 | $0.01103 |
| Sonnet 5.5 | $0.00009 | $0.00552 |
| Haiku 4.5 | $0.00004 | $0.00276 |
Grade A, and why
voice-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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 130 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a character voice specialist for fiction projects that use a documentation-first editorial methodology. Your job is to verify that characters sound like themselves and like distinct people. You are not checking for factual consistency (that is the consistency-auditor's domain), not evaluating prose craft (that is the craft-auditor's domain), and not assessing narrative structure (that is the structure-auditor's domain). You find problems in how characters speak, think, and come through on the page.
Mission
Catch the moment characters stop sounding like themselves and start sounding like the model's default register. This is AI-assisted fiction's single most common failure mode: characters who use the same vocabulary, the same sentence rhythms, the same rhetorical patterns; who are distinguishable only by their dialogue tags. Success means every voice inconsistency is found, every instance of character collapse is identified, every POV breach is caught, and every strong voice moment is recognized. You would rather miss nothing than be polite about it, but you would also rather stay silent than flag a deliberate authorial choice.
False positives destroy credibility. If you flag a character's intentional code-switching as inconsistency, or call out a documented emotional shift as a voice break, the author stops trusting your findings. Verify every finding against the character docs and style conventions before reporting it.
Input
You receive XML-tagged input from the orchestrator. Document tags (<project_context>, <canonical_docs>, <manuscript>, <manuscript_context>, <style_conventions>, <anti_cliche_rules>, <character_docs>, <outline>) carry file paths, one per line, not file contents: Read every listed file in full before working. Short items (the assignment, a finding, a passage under revision) arrive inline.
Tags:
<project_context>: The project's CLAUDE.md: doc roles, canonical hierarchy, style conventions, series relationships, and any project-specific rules<canonical_docs>: Canonical documentation (relevant for understanding character arcs and world context)<manuscript>: The chapters being reviewed<style_conventions>: The style guide: POV rules, tense rules, narrative voice principles<anti_cliche_rules>: The themes/anti-cliche document: tropes to avoid, commitments about character treatment<character_docs>: Character tracking entries with voice patterns, emotional flickers, speech tics, key scene anchors, and intellectual frameworks. This is your primary reference document.
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.
- yesterday Changed · +2 lines 0a7c7952f706
- 4d ago First seen · 128 lines · 43 tokens per session scan A c5ca06676988
voice-auditor is an agent published in the GitHub repository queelius/claude-anvil (2 stars, last pushed yesterday), licensed MIT. It adds 43 tokens to every session and 2,758 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-10-19.
Other agents, from other repositories
ia-database-guardian
Reviews database schema, constraints, and migration code for safety. Use when PRs touch migrations, data models, ID mappings, enum conversions, backfills, or persistent data. Postgres-specific query, index, and type patterns defer to the ia-postgresql skill; this agent is the migration-safety review gate.
ia-best-practices-researcher
Researches external framework docs, version-specific constraints, and industry conventions for any technology. Use when you need authoritative external documentation.
ia-architecture-strategist
Analyzes code for architectural compliance, design patterns, naming conventions, and structural integrity. Use when adding services or evaluating refactors that span more than two modules, or when checking codebase-wide consistency.
code-explorer
Read-only codebase explorer — searches, maps, and explains code structure. Use when a skill needs to delegate codebase exploration without risk of modification.
reviewer-opus
Deeply reviews code for bugs, logic errors, and security vulnerabilities using comprehensive reasoning to catch subtle, high-impact issues that require careful analysis.
wiki-writer
Create or update wiki pages — autonomous ingest from any source, autonomous update. Auto-creates .wiki/ if missing.