PenguinHarness is a local-first platform in which multiple AI agents create, evaluate, optimize, and deploy agent applications. It is for people building AI software who want agents to generate applications and improve their own behavior through skills.
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.
npx agentmods add skills/prism-shadow/penguin-harness/humanizernpx skills add Prism-Shadow/penguin-harness --skill humanizergit clone --depth 1 https://github.com/Prism-Shadow/penguin-harnessWrote 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/skills/prism-shadow/penguin-harness/humanizer)<a href="https://agentmods.dev/skills/prism-shadow/penguin-harness/humanizer"><img src="https://agentmods.dev/badge/skills/prism-shadow/penguin-harness/humanizer.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 | $0.00136 | $0.01538 |
| Opus 5 | $0.00068 | $0.00769 |
| Sonnet 5 | $0.00027 | $0.00308 |
| Haiku 4.5 | $0.00014 | $0.00154 |
Grade A, and why
humanizer 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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Humanizer
Make prose read like edited human writing: the register of books, quality newspapers and encyclopedia entries. Everything here was derived empirically, across seven rounds of drafting, blind editorial review and revision, documented with counts in reference/case-study.md. The working surface is deliberately small: a writer boxed in by a long checklist produces compliance, not prose — that failure mode is the case study's best-documented finding. Draft with the principles below; diagnose with the catalog afterwards.
Before you start
If the invocation carries no text and no assignment, ask for one: the draft to edit, or the topic, length, audience and language to write fresh. Settle two things early. The register: books, newspapers and encyclopedias are the default target, while marketing, speeches and reference documentation legitimately bend these rules, so confirm how far to go. And any hard length target: humanizing shrinks text, and gaps are filled with substance, never padding. If the genre needs material nobody has gathered — reportage needs a scene, a person, a quotation — say so and get it, or agree to relabel the piece; never fake the texture.
Core principles
- No pattern twice. Whatever the figure — a contrast frame, a triad, a cleft, an opener shape, a paragraph arc, a metaphor, even a favorite connective particle — its second consecutive use is a rhythm and its third is a stencil. Vary sentence length, clause weight, paragraph attack and closer; if every paragraph advances by the same move, swap engines somewhere.
- Density is anchored facts in whole grammar. Names, dates, numbers, mechanisms and worked examples carry the argument; hype, era openers, phantom crowds ("faster than most expected" — who?) and concepts pushing concepts carry nothing. Compress by dropping padding, not grammar: subjects stay, first mentions get their full noun, and the event that matters gets a sentence of its own.
- Cap the quotables. One or two turned phrases can carry a piece, counting every shape: chiasmus, mirrored re-description, balanced antithesis, aphoristic kickers. State the thesis once and develop it with new material or cut the echo. Most paragraphs end flat, and an argument survives a dud.
- A real writer, with the material. Opinion owns its anecdotes and risks a judgment of its own; reportage has stood somewhere; reference registers keep the writer invisible without chaperoning the reader. Claims sized to the evidence, honest hedges kept, references anchored (who, where, when), and one non-obvious source beats a second canonical one.
- Structure serves content. Open on ground, not on a cold verdict and not on an era; paragraphs develop one idea across several sentences; headings and bullets appear only where content is genuinely enumerable; the piece ends where the information ends.
- Write from inside the language. Native idiom — a sentence that back-translates cleanly into another language was composed there, so recompose it — native punctuation at native frequency (a dash where the register expects one beats zero), and the venue's typography held consistently. Per-language budgets and surface forms:
reference/language-cues.md, which indexes onereference/<lang>-cues.mdfile per language. - True, verified, calibrated. Never invent specifics. Re-derive every mechanism example from the stated mechanism, split fused attributions, and verify or delete every superlative. One fluent falsehood outweighs any amount of style.
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 · -5 lines 5e5346ffe6f8
- 6d ago First seen · 45 lines · 136 tokens per session scan A bef2d8c650f4
humanizer is a skill published in the GitHub repository Prism-Shadow/penguin-harness (1,991 stars, last pushed today), licensed Apache-2.0. It adds 136 tokens to every session and 1,538 once invoked, about $0.0007 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 2026-08-30.
Other skills, from other repositories
harness-creator
Build, audit, and improve harnesses that make AI coding agents reliable: AGENTS.md/CLAUDE.md instruction files, feature/state tracking, verification gates, scope boundaries, session handoff, memory persistence, context budgets, tool-permission safety, and multi-agent coordination. Use this whenever a coding agent is…
agentfield-use
Whenever you have a discrete task to perform — one the user delegated, or one that arose inside your own work — check FIRST whether an installed AgentField agent covers it, and offload to it by default when one does. Coverage, not task size, is the test: even a small job goes to a covering agent. The check is cheap …
agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a…
browser-workflow-skill-builder
Create robust browser automation skills for sites like LinkedIn, X/Twitter, YouTube, Stripe, or other logged-in web apps by saving page context, analyzing HTML/CSS, writing skill-local JS extract/verify scripts, and using CSS selector actions with screenshot verification.
cli-skill-design
Design a co CLI surface and its SKILL.md together so an agent can drive it without guessing — every command ends by naming the next one, --help lists everything, and every failure says what to run instead. Use when adding a new CLI command group, writing or rewriting a SKILL.md for one, or auditing an existing one.
oo
Connect to a remote ConnectOnion agent, set up a publishable identity, publish selected skills, or follow and sync another publisher. Use for a 0x agent address, /oo, co announce, or co sub workflows; not for ordinary local-agent tasks.