{
 "schema_version": 1,
 "generated": "2026-10-19",
 "id": "google-research-harness2--benchmark-agent-harness-without-answer-leaks",
 "repo": "google-research/harness2",
 "repo_url": "https://github.com/google-research/harness2",
 "repo_slug": "google-research-harness2",
 "title": "Run a clean benchmark with network limits against rubric leaks",
 "page_title": "How to run a clean benchmark with network limits against rubric leaks with harness2",
 "summary": "Lock down network access before evaluating so an agent cannot look up public benchmark answers.",
 "who": "Researcher reporting harness-improvement numbers who must show the agent could not read public rubrics",
 "problem": "Executors have open web access and public rubrics may exist online, so scores can be inflated by looking up answers.",
 "outcome": "A benchmark run done under network restrictions, giving numbers you can defend as not contaminated by leaked grading criteria.",
 "steps": [
  "Read the web access section of docs/EXPERIMENTS.md and apply its network restrictions.",
  "Preview task selection with no model calls by adding --dry-run to your harness2 command.",
  "Run the real refinement: harness2 --bench lab --k 1 --mode parallel --only-domains antitrust-competition --max-tasks 1",
  "Judge: harness2-judge --bench lab --k 1 --mode parallel",
  "Report: harness2-report --bench lab --k 1 --mode parallel"
 ],
 "agent_prompt": "Before running any model-backed benchmark in google-research/harness2, read the web access section in docs/EXPERIMENTS.md and tell me what network restrictions to configure for <benchmark>. Then run a --dry-run preview and verify task selection looks right.",
 "agents": [
  "any"
 ],
 "difficulty": "Advanced",
 "time": "1 hour",
 "basis": "field",
 "caveat": "Authorized use only: run this only inside infrastructure you control, since the README states executors have unrestricted network access and Codex runs with its sandbox disabled, and public rubrics may be online.",
 "category": "Research & Knowledge",
 "tags": [
  "Research & Knowledge"
 ],
 "lang": "python",
 "requirements": [
  "Linux, Python 3.12 and uv for installation",
  "No credentials needed for the local demo (harness2-demo)",
  "Model-backed runs need Node.js 20+, the Google Cloud CLI, a Vertex AI project and model IDs in .env, and gcloud application-default login",
  "LAB needs Pandoc and Poppler; WorkBuddy-Bench needs Docker",
  "Model-backed benchmark runs incur API costs"
 ],
 "risk": "dual-use",
 "seen_in_the_wild": true,
 "best": false,
 "score": 0.532,
 "evidence": [
  {
   "platform": "x",
   "url": "https://x.com/i/web/status/2112265188091265326",
   "posted": "2026-10-19"
  },
  {
   "platform": "x",
   "url": "https://x.com/i/web/status/2112254069184344121",
   "posted": "2026-10-19"
  }
 ],
 "url": "https://dedupedai.com/use-cases/google-research-harness2/benchmark-agent-harness-without-answer-leaks/",
 "md": "https://dedupedai.com/use-cases/google-research-harness2/benchmark-agent-harness-without-answer-leaks/index.md",
 "disclaimer": "Deduped lists open-source repositories that belong to other people. We do not own, maintain, audit, or vouch for any of them, and a repo appearing here is not an endorsement or a safety check. The use cases, steps, and prompts are written by AI and may be incomplete, outdated, or just wrong. Where a use case is marked \"Seen in the wild\", we link to public posts where someone described something similar, but we have not verified those posts and the use case itself is still our AI's own write-up. Use all of this at your own risk. Only run code, commands, or agents on systems and accounts you are authorized to use, and read the code and every command before you run it, especially anything that touches credentials, files, or the network. Nothing here is professional, security, legal, or financial advice. All trademarks and project names belong to their respective owners, and their use here does not imply affiliation. Content is provided as is, without warranties of any kind. Found a mistake, a safety problem, or something you want removed? Email hello@dedupedai.com and we will look at it promptly."
}