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/ofershap/ai-humanizerWrote 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/rules/ofershap/ai-humanizer/best-practices)<a href="https://agentmods.dev/rules/ofershap/ai-humanizer/best-practices"><img src="https://agentmods.dev/badge/rules/ofershap/ai-humanizer/best-practices/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/rules/ofershap/ai-humanizer/best-practices"><img src="https://agentmods.dev/badge/rules/ofershap/ai-humanizer/best-practices.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 | $0.00349 | $0.00349 |
| Opus 5.5 | $0.00140 | $0.00140 |
| Sonnet 5.5 | $0.00070 | $0.00070 |
| Haiku 4.5 | $0.00035 | $0.00035 |
Grade A, and why
best-practices 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 1mo 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.
What it actually says
- Never use em dashes (—) as separators. Use hyphens with spaces ( - ), periods, or restructure the sentence.
- Never use these words: delve, tapestry, leverage, utilize, facilitate, streamline, paramount, pivotal, meticulous, holistic, robust, comprehensive, multifaceted, harness, foster, bolster, seamless, cutting-edge, game-changing, revolutionary, groundbreaking, nuanced, intricate, elevate, empower, unleash.
- Never open with "In today's...", "It's worth noting...", "Let me explain...", "At its core..." or close with "In summary...", "Happy coding!", "In conclusion...".
- Never add narrating code comments that describe what the code does. Only comment non-obvious intent, constraints, or trade-offs.
- Never add git co-author trailers for AI tools (Co-authored-by: Cursor, Copilot, etc.) or "Generated by" attribution.
- Vary sentence length. Mix short direct statements with longer ones. Never write 4+ sentences of similar length in a row.
- Use plain words. "important" not "paramount", "careful" not "meticulous", "use" not "leverage", "improve" not "optimize".
- Keep commit messages short and lowercase. No marketing language in commit bodies.
- Never over-format: no bold-label bullets (Feature: desc), no uniform list lengths, no tricolons in every list.
- Write READMEs and docs in a direct, sparse style. State what the thing does, not why it's amazing.
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.
- 1mo ago First seen · 16 lines · 349 tokens per session scan A 2d58b2743428
best-practices is a cursor rule published in the GitHub repository ofershap/ai-humanizer (5 stars, last pushed 8mo ago), licensed MIT. It adds 349 tokens to every session, about $0.0014 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 2026-08-31.
Other cursor rules, from other repositories
humanize-text-skill
Audit and rewrite Chinese or English content to remove AI tone, then pull it toward a target human voice. Activate when editing prose-heavy files such as Markdown, docs, blog posts, READMEs, release notes, emails, issue replies, or social copy.
deai
A rule set for editing Chinese Markdown manuscripts sentence by sentence or paragraph by paragraph to reduce signs of AI-written text.
checks
How to add or modify a scanner check.
unslop
Humanize assistant output. Drop AI-isms, engineer burstiness, preserve technical accuracy.
arcjet-ai-apps
Security guidance for AI/LLM endpoints — layered Arcjet protection with prompt injection, PII blocking, and token budgets.
arcjet-express
Arcjet patterns for Express, Node.js, Fastify, Hono, NestJS, Bun, and Deno — correct adapters, no middleware, proxy config.