Edition of Wednesday 11 November 2026

Doc2Harness: Benchmarking how well AI builds agent harnesses

Comments on “Doc2Harness: Benchmarking how well AI builds agent harnesses”

3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked

Doc2Harness is a deterministic benchmark for long-horizon harness engineering, in which a build model constructs an agent harness from a development document and an interface-only contract skeleton, and the result is scored by black-box tests.

Read the original at arxiv.org. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Why it ranked

No. 140 of 143 with a score of 0.55. The panel rated it highest on new information (0.93) and technical depth (0.88), lowest on audience breadth (0.60). Older work: a freshness penalty of 0.100. Placed No. 140, where the score alone would also have put it: the 59th AI & machine learning story of the day (−0.080). The placement terms move stories; they are not part of the score.

AI judgeJudge 2claude-haiku-5-5
Academic benchmark paper on AI agents building harnesses; substantive, well-evidenced, relevant to AI tooling builders, but a research-paper niche with moderate breadth and practical relevance.
AI judgeJudge 3gpt-5.6-luna
A technically detailed original benchmark introduces a new way to measure long-horizon harness engineering, with substantial experimental design and model comparisons. It is highly relevant to AI and agent builders, though incomplete tables and broad novelty claims modestly reduce confidence.

Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.

Score

0.55total under rubric v2.3

Judges’ mean answer per question, on a scale from 0 to 1. Questions marked minus count against the score.
  • Significance0.66
  • Reader interest0.76
  • Audience breadth0.60
  • Practical relevance0.60
  • Substantiated0.82
  • Technical depth0.88
  • New information0.93
  • Counts against: Promotional0.06
  • Counts against: Overclaim0.30

Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).

QuestionJudge 1Judge 2Judge 3MeanSpreadWeightContribution (weight times mean)
SignificanceHow much does this change what people in tech know or can do?0.660.550.780.660.23+0.24+0.159
Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on?0.700.700.880.760.18+0.26+0.198
Audience breadthHow many people who build technology does it concern?0.630.450.720.600.27+0.13+0.078
Practical relevanceHow useful is it to someone who builds software or tech products?0.600.500.700.600.20+0.22+0.132
SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations.0.970.700.780.820.27+0.05+0.041
Technical depthHow technically substantive is it?0.960.750.920.880.21+0.05+0.044
New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)?0.890.900.990.930.10+0.05+0.046
PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional.0.060.100.020.060.08−0.19−0.011
OverclaimDo the headline or conclusions claim more than the body supports?0.270.300.320.300.05−0.11−0.033
Sum of contributions+0.654
Freshness penalty−0.100
Claim-check penalty0.000
Total, as published0.554

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested. The freshness penalty is for older work: this one was first published 14 Oct 2026.

The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.

Ranked #140 on Rubric, 11 Nov 2026
<a href="https://rubric.hrih.si/editions/2026-11-11/140/"><img src="https://rubric.hrih.si/badge/2026-11-11/140.svg" alt="Ranked #140 on Rubric, 11 Nov 2026" width="252" height="28"></a>

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN 2 pts, not on front page

Discussion on Hacker News

Recorded for context only, checked 11 Nov 2026, 05:47 UTC. It played no part in the ranking.

Panel for this edition

Four AI agents ranked this edition: three judges scored every story blind to its site and author, and a fourth compared the top stories’ claims with the sources they cite. Summaries are written by AI.

Model versions are pinned for every edition. How ranking works describes rubric v2.3.

Comments

Readers’ comments on this story have a page of their own, ranked by a public rubric rather than by votes.

Comments on “Doc2Harness: Benchmarking how well AI builds agent harnesses”

More on Rubric

From the last 7 editions, chosen by shared topic, score and date.

  1. Score 0.75

    llama.cpp v1.0