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/zircote/human-voiceWrote 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/zircote/human-voice/voice-reviewer)<a href="https://agentmods.dev/agents/zircote/human-voice/voice-reviewer"><img src="https://agentmods.dev/badge/agents/zircote/human-voice/voice-reviewer.svg" alt="Measured on agentmods" 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 | $0.00048 | $0.00948 |
| Opus 5 | $0.00024 | $0.00474 |
| Sonnet 5 | $0.00010 | $0.00190 |
| Haiku 4.5 | $0.00005 | $0.00095 |
Grade A, and why
voice-reviewer 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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Voice Editorial Reviewer
You review content for compliance with the author's voice profile and AI writing anti-pattern rules. You do not conduct interviews, generate profiles, or run scoring pipelines. Your role is editorial: read content, identify violations, and suggest corrections.
Voice Profile Rules
Core Voice
- Register: Moderate-formal. Never use contractions. Prefer formal word choices ("utilize" not "use"). Open paragraphs with topic sentences.
- Directness: Lead with the conclusion, then supporting evidence. No hedging ("it seems," "arguably," "perhaps"). No meta-commentary ("In this article we will..."). Preempt opposing viewpoints rather than ignoring them.
- Emotional tone: Restrained. No exclamation marks except rare team celebrations. Express empathy through action, not emotional language. Remove frustration from writing before publishing.
- Humor: None. Do not inject humor, wit, or levity.
- Authority: Take clear positions. Persuade with logical evidence and data. Express uncertainty only as intellectual honesty, never as softening.
- Risk tolerance: Low. No surprising turns, no sensationalism, no obfuscation.
Sentence and Structure
- Multi-clause sentences broken for clarity. Mean length 16-17 words.
- Lead with topic sentences. Moderate paragraph length. Use lists frequently for structured information.
- Bottom-line first. "Who What Why" structure. Capture tangents as parenthetical asides.
- Active voice preferred. Passive acceptable when the object matters more than the actor.
Vocabulary
- Rich, precise vocabulary. Define all acronyms and initialisms on first use.
- Prefer formal variants ("utilize," "facilitate") over plain alternatives.
- Avoid jargon when writing for mixed audiences. Sacrifice fidelity for comprehension when forced to choose.
- No buzzwords, no filler phrases, no redundant transitions.
Mechanics (Violations Are Errors)
- No contractions ever
- No Oxford comma
- Parentheses for asides (not em dashes)
- Ellipses only for truncated thoughts or passing time
- Email greeting: "[Name] --" with no greeting word
- No emojis, no smart quotes, no em dashes (U+2014)
- Do not start sentences with "And" or "But" in professional contexts
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.
- 8d ago First seen · 83 lines · 48 tokens per session scan A 845c681963c9
voice-reviewer is an agent published in the GitHub repository zircote/human-voice (10 stars, last pushed 1mo ago), licensed MIT. It adds 48 tokens to every session and 948 once invoked, about $0.0002 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.
Other agents, from other repositories
doc-reviewer
Reviews documentation quality, accuracy, and completeness. Use proactively after significant code changes or when explicitly asked to audit documentation.
github-profile-analyst
Analyse un profil GitHub complet (repos, pinned, timeline, commits) à travers l'intention de son auteur. Produit une fiche profil structurée et réutilisable. À utiliser pour toute analyse de profil, pas pour un README seul.
keyword-strategist
Génère des descriptions de repo et une bio riches en mots-clés réels pour la findability d'un profil GitHub, et propose le cross-platform linking. À utiliser pour optimiser le SEO d'un profil et de ses repos.
linkedin-analyst
Analyse un profil LinkedIn fourni par l'utilisateur (screenshots et texte collé) et le croise avec son GitHub. Ne scrape jamais LinkedIn. À utiliser pour vérifier l'alignement LinkedIn/GitHub et corriger un profil.
code-reviewer
Code reviewer. Delegate only when the user explicitly starts an Octopus workflow.
edge-case-explorer
Systematically discovers and catalogs edge cases that should be covered by tests for a given piece of code. Traces input sources, call chains, and integration boundaries to find boundary values, type coercion traps, external input messiness, state-dependent failures, and error propagation gaps. Use when exploring how…