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 instructions/steel-dev/leaderboard/agents-mdgit clone --depth 1 https://github.com/steel-dev/leaderboardWrote 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/instructions/steel-dev/leaderboard/agents-md)<a href="https://agentmods.dev/instructions/steel-dev/leaderboard/agents-md"><img src="https://agentmods.dev/badge/instructions/steel-dev/leaderboard/agents-md.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 | $0.00630 | $0.00630 |
| Opus 5 | $0.00315 | $0.00315 |
| Sonnet 5 | $0.00126 | $0.00126 |
| Haiku 4.5 | $0.00063 | $0.00063 |
Grade A, and why
leaderboard AGENTS.md 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 3d 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 — 28 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Guidelines
Project Structure & Module Organization
This is an Astro + TypeScript site for the Steel browser-agent leaderboard. The main page is src/pages/index.astro. Markdown-style endpoints index.md.ts and llms-full.txt.ts expose leaderboard data for tools and LLMs. Shared UI lives in src/components/, page shells in src/layouts/, and leaderboard data in src/lib/leaderboard.ts. Static assets belong in public/ or src/assets/. Research notes may live under docs/research/<slug>/.
Build, Test, and Development Commands
Use npm install to set up dependencies. Key scripts:
npm run dev: start the Astro dev server.npm run build: produce the production build indist/.npm run preview: serve the built site locally.npm run lint: run ESLint across the repo.npm run lint:fix: apply autofixes where possible.npm run update-readme: rebuild the README leaderboard table fromsrc/lib/leaderboard.ts.npm run discover: query arxiv for recent benchmark papers and write candidates to.discovery/run.json(gitignored). Used by thebenchmark-discoveryskill in.claude/skills/.
Maintainer skills
The .claude/skills/benchmark-discovery/ skill drives the discovery flow. Invoke it from a Claude Code session to do a sweep, classify candidates (proposed / variant / methodology), and optionally append the result to the rolling discovery-labeled issue. The skill never writes to src/data/ — maintainers verify candidates manually before merging.
Coding Style & Naming Conventions
Follow the existing Prettier config: 2-space indentation, semicolons, double quotes, printWidth: 100, and trailing commas where valid in ES5. ESLint enforces no unused variables and disallows any in TypeScript. Keep Astro components and layouts in PascalCase (for example, LeaderboardTable.astro), use camelCase for exported helpers, and keep slug-like files and folders in kebab-case (for example, docs/research/webvoyager/).
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.
- 3d ago First seen · 28 lines · 630 tokens per session scan A 98317befb60e
leaderboard AGENTS.md is an instructions file published in the GitHub repository steel-dev/leaderboard (47 stars, last pushed 8d ago), licensed MIT. It adds 630 tokens to every session, about $0.0032 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 instructions, from other repositories
obscura AGENTS.md
Instructions for h4ckf0r0day/obscura, covering agents.md, build, rendering and stealth, no rendering, with rustls or stealth and test.
stealth-browser-mcp AGENTS.md
Instructions for vibheksoni/stealth-browser-mcp, covering agent instructions, start here, recommended proxy services, environment and validation.
playwright-rust CLAUDE.md
Instructions for padamson/playwright-rust, covering claude.md, project, repository layout, skills (procedural reference) and documentation hierarchy.
pi AGENTS.md
Instructions for TGYD-helige/pi, covering pi extensions monorepo — agent guidelines, project overview, required agent skills, ponytail — minimal-code discipline and mattpocock/skills — engineering workflow skills.
flyto-core CLAUDE.md
Instructions for flytohub/flyto-core, covering claude notes, cross-agent handoff and shared code intelligence.
mcp-stealth-chrome AGENTS.md
Instructions for RobithYusuf/mcp-stealth-chrome, covering mcp-stealth-chrome — agent guide, product and repository map, hard constraints, working defaults and architecture invariants.